Amyloid precursor protein (APP), an integral membrane protein with unknown function, is the precursor of a series of soluble proteins and peptides, including the A4 amyloid peptide ((1, 2, 3, 4, 5) ; reviewed in (6) ). This peptide is the main component of the amyloid plaques, one of the neuropathological hallmarks of Alzheimer's Disease (AD)(7, 8) . Although the direct relationship between A4 amyloid peptide deposition and the progressive neuronal death observed in the brains of these patients remains to be clarified, genetic data demonstrate the central role of APP in the pathogenesis of AD, since point mutations in APP are linked to the familial form of AD(9, 10, 11, 12, 13) . Furthermore, a gene dosage effect of the APP gene on chromosome 21 is implicated in AD associated with trisomy 21 (Down's syndrome)(14) . Finally, the A4 peptide is under certain experimental conditions toxic for neuronal cells in vitro(15, 16) .

Proteolytic processing of APP occurs at three sites in the protein, either amino-terminal and carboxyl-terminal of the A4 sequence, producing the A4 peptide, or in the middle of the A4 peptide to release the ectodomain and thus precluding amyloidogenesis(6, 17) . The ``amyloidogenic'' cleavages are mediated by hypothetical - and -secretases, whereas the ``nonamyloidogenic'' cleavage is mediated by -secretase(s)(17, 18) . Although none of these enzymes have been identified yet, available evidence and conjecture points to some unusual characteristics. The -secretase has to cleave in the APP transmembrane region, raising conceptual problems regarding the mechanism. The -secretase has an extremely relaxed specificity regarding the primary amino acid sequence cleaved but requires the presence of the transmembrane domain of APP(19, 20, 21) . It was proposed that several proteinases can perform -secretase cleavage(18) , which in turn raises the question of how other integral membrane proteins avoid this activity. An ``secretase compartment'' to which APP is specifically targeted was proposed to be the trans-Golgi network (TGN) or the transport vesicles en route toward the cell surface(21, 22, 23, 24) . Since deletion of the cytoplasmic domain carrying the putative sorting signals GY and NPTY increased the secretion of APP, these signals were suggested to target APP toward nonsecretory, intracellular pathways (21) . The different intracellular pathways that are open for APP, characterized by specific processing and sorting steps in specific subcellular compartments, have to be studied. Given the complexity of the cellular protein sorting mechanisms, it is conceivable that the different genetic and epigenetic factors believed to underlie the pathogenesis of AD (25, 26, 27) interfere at one or another level with the mechanisms governing the cellular processing and sorting pathways followed by APP. In that regard, it should be noted that the amyloid lesions of AD are associated with two types of polarized cells, i.e. neurons and endothelial cells.

The best characterized model to study protein sorting is the Madin-Darby canine kidney (MDCK) cell line(28, 29, 30) . These cells form a well polarized, tight epithelial cell layer in vitro, separating apical and basolateral compartments. Moreover, an interesting parallel between MDCK cells and neuronal cells is the fact that proteins of the apical and basolateral membrane of MDCK cells are targeted to axonal and somatodendritic processes, respectively, in primary cultures of hippocampal neurons(31, 32, 33) .

To extend our previous studies of APP processing and secretion in unpolarized cells, we decided to approach the problem of polarized sorting and metabolism of APP in the MDCK model. Previously we have demonstrated that APP secretion is subject to modulation by metabolic factors and structural determinants(21) . We now report that APP is secreted in MDCK cells in a strictly polarized fashion, confirming a recent report(34) . The correct sorting of wild type, soluble APP is extremely sensitive to pH changes in intracellular compartments as demonstrated with primary amines and bafilomycins, indicating the involvement of an acid-sensitive sorting mechanism. We further demonstrate that potential basolateral sorting signals in the cytoplasmic domain of APP do not determine the polarized secretion of APP and that the major basolateral sorting signal is located in the extracellular domain. Furthermore, we find that a fraction of APP terminating at either the - or the -secretase cleavage site is missorted to the apical compartment. We finally demonstrate that APP containing the ``swedish'' mutation(12) , which is linked to familial AD, not only causes overproduction of A4 peptide in certain cell types (35, 36) but is also partially missorted after -secretase cleavage to the apical compartment in polarized MDCK cells.

EXPERIMENTAL PROCEDURES

Materials

Rabbit antiserum B2/3 against mouse APP was previously characterized (21, 22) ; goat antiserum 207 against soluble APP (37) was generously provided by Dr. B. Greenberg (Cephalon, West Chester, PA); rabbit antiserum R1736 against residues 2-15 of the A4 peptide (4, 34) was kindly provided by D. Selkoe (Boston).

The cDNA for mouse APP has previously been characterized(38) . CDNA coding for human APP695 was kindly provided by Dr. R. Scott (Cephalon, West Chester, PA). CDNA coding for human APP695 containing the early onset familial Alzheimer's Disease (EOFAD) mutations APP695(H)617A:G (13) , APP695(H)618E:Q(9) , APP695(H)595K:N/596 M:L, swedish mutation(12) , and APP695(H)642V:I (11) were synthesized by site-directed mutagenesis (L. Hendriks and C. Van Broeckhoven, Antwerp, Belgium).

Cell Culture

The tightness of cell layers was tested by adding [H]inulin (Amersham Corp.) to the apical compartment and measuring the amount of label that leaked to the basolateral compartment after 1 h. This was always less than 1%. The polarity of the cells in the culture conditions used was ascertained by confirming the polarized secretion of gp80 (39) and the polarized uptake of [S]methionine. In addition, preliminary experiments proved that the secretion of endogenous APP751/770 was actually a very reliable basolateral marker, which functions in all transfection experiments as an internal control (see ``Results'').

For transfection, 30 µg of linearized pRC/RSV plasmid containing the cDNA of mouse APP695 or mutants thereof or 30 µg of circular pSG5 plasmid containing the EOFAD APP695(H) mutants with 1 µg of circular pRC/RSV plasmid was added to 2.5 10 cells suspended in 0.5 ml of phosphate-buffered saline. Electroporation was done at 260 V, 960 microfarads (Gene-Pulser; Bio-Rad). The time constant varied between 16 and 22 ms. After transfection, cells were cultured for 24-48 h in medium supplemented with penicillin/streptomycin. Selection of transfected cells was performed in medium containing 700 µg/ml G418 sulfate (Geneticin; Life Technologies, Inc.). After 2 weeks, cells were tested for expression of the transfected constructs. Transfected cells were used for a maximum of five passages. To demonstrate that a high level of overexpressed, transfected APP per se is not sufficient to target APP toward the apical compartment, high expression clones of MDCK cells transfected with APP695(M) and APP695(M)Ala666*, were isolated by limiting dilution.

Concentrated stocks of drugs were diluted into culture medium. For the alkalization experiments, cell layers were preincubated with the drugs for 30 min, and labeling was performed in the presence of the drug. Phorbol esters (PMA and Pdbu) and cholera toxin, pertussis toxin, and forskolin were present only during labeling. Control experiments ascertained that dimethyl sulfoxide used as a solvent for certain drugs at the dilutions used (1:1000) did not affect the cells.

cDNA Constructs

cDNA coding for human APP695(H) was cloned as a HindIII fragment in the expression vector pSG5 (Stratagene). This vector allows transcription of the cloned cDNA from the SV40 early promoter. Mutants of mouse and human APP695 (see Fig. 1) were generated by site-directed mutagenesis in the pSG5 plasmid(40) . Codons coding for Tyr 653, Leu613, and Asp597, respectively, were mutated toward a TAG stop codon, yielding APP695(M)Y653*, APP695(M)L613*, and APP695(M)D597* (see Fig. 1). APP695(M)A626* was made by introducing a cassette in the BbeI restriction site of APP(M)695, changing codon 626 to a stop codon. The generation of APP695(M)A666*, previously named APP695TRUNC, has been described in detail elsewhere(21) .

Figure 1: Deletion mutants of mouse APP695. The domain structure for mouse APP (which is 97.6% identical to human APP) (38) is shown. SP, signal peptide; A4, amyloid peptide; TM, transmembrane domain; CD, cytoplasmic domain. The -secretase cleavage site is indicated by a solid triangle. The displayed deletion mutants were obtained by introducing stop codons at the indicated positions (see ``Experimental Procedures''). Below the structure of APP695(M), the regions to which antibodies 207, B2/3, and R 1736 were raised are indicated.

All constructs were analyzed by restriction analysis and by sequencing of the mutated sites.

Metabolic Labeling, Double Immune Precipitation, and Quantitation

In double immune precipitation assays, 1 ml of basolateral medium or 0.6 ml of apical medium was used. Immune precipitation was performed with 20 µl of rabbit antiserum B2/3 against mouse APP, 3 µl of goat antiserum 207 against soluble APP, 3 µl of R1736 prepared against synthetic human A4 residues 2-15, 20 µl of rabbit antiserum B7/7, or 20 µl of antiserum SGY2134 prepared against synthetic A4 peptide. All samples were brought to equal volumes with phosphate-buffered saline and 10 double immune precipitation buffer (Tris-buffered saline containing 10% Triton X-100, 10% sodium deoxycholate, and 1% SDS(41) ). Recovery of the immune complexes was done with protein A-Sepharose CL4B (Pharmacia Biotech Inc.) or protein G-agarose (Immunopure; Pierce). Processing of immune complexes was essentially as described previously(21, 41) . 6% polyacrylamide gels were impregnated with dimethyl sulfoxide/2,5-diphenyloxazole, dried, and exposed to preflashed Hyperfilm MP (Amersham Corp.)(21) . Exposure varied from 1 day to 3 weeks. Quantitation was performed by densitometric laser scanning(21) .

RESULTS

Basolateral Secretion of Endogenous and Transfected APP in Polarized MDCK Cells

Figure 2: Basolateral secretion of APP in polarized MDCK cells. Panel a, endogenous APP was immune precipitated with antibody 207 from apical (A) and basolateral culture media (B) of untransfected MDCK cells after a 2-h metabolic labeling with [S]methionine. Negative controls, using only protein G-sepharose are shown on the right. Note the presence of a protein of about 500 kDa in the apical compartment, precipitated by protein G. Panel b, MDCK cells were transfected with mouse APP695 cDNA. A high expressing clone was obtained by limiting dilution. Cells were labeled for 2 h, and immune precipitation was performed with antibody B2/3. Lanes on the right are negative controls using MDCK cells transfected with pRC/RSV vector alone. Panel c, transfected mouse APP695 and endogenous canine APP751/770 were immune precipitated with antibody B2/3. Cells were pulse-labeled for 10 min and chased for the time periods indicated. A, apical, B, basolateral compartment. Soluble mouse APP695 as well as canine APP751/770 is weakly visible after 20 min of chase in the basolateral compartment, whereas no APP is precipitated from the apical compartment, even after 180 min of chase.

MDCK cells were transfected with recombinant DNA constructs coding for mouse or human APP695 or for human APP770. Transfected APP695 yielded proteins that migrated with a higher mobility in SDS-PAGE than endogenous APP (Fig. 2, b and c), whereas transfected APP770 co-migrated with the endogenous protein (result not shown), demonstrating that the endogenously produced canine APP contained the Kunitz type proteinase inhibitor domain. We could not determine whether either APP751 or APP770 or both were present, which is without further consequences for our experiments. Overexposure of immunoblots or immune precipitations indicated that the MDCK cells also produced very small amounts of endogenous APP695. The basolateral secretion of endogenous canine APP was used in subsequent transfection experiments as an internal control, demonstrating that the sorting machinery remained operational in the transfected cells (Fig. 2c and all other figures).

When transfected mouse APP695 was expressed at a level that was at least 10-fold higher than that of endogenous APP, secretion remained almost exclusively into the basolateral compartment (Fig. 2b), which indicated that the APP-sorting mechanism is characterized by a high efficiency and a high capacity. In addition, pulse-chase experiments demonstrated that endogenous and transfected APP were secreted with similar fast kinetics into the basolateral compartment (Fig. 2c). This demonstrated that endogenous and transfected APP were transported directly to the basolateral side and that secretion of APP does not involve transcytosis. After a 3-h chase, no APP was detected in the apical compartment, confirming the strongly polarized secretion of APP (Fig. 2c).

Alkalization of Internal Compartments Randomizes Secretion of APP

Figure 3: Effect of methylamine and pH on secretion of endogenous APP in MDCK cells. Panel a, untransfected MDCK cells were preincubated for 30 min and labeled for 120 min in culture medium buffered at the indicated pH, without (-MA) or with the addition of 20 mM methylamine (+MA). APP was immune precipitated with antibody 207. Panel b, quantitative analysis of three independent experiments performed as in panel a. The increased apical (A) and decreased basolateral (B) secretion of APP in the presence of methylamine (+MA), compared with the basolateral secretion of APP in the absence (-MA) of methylamine is demonstrated. Changing the pH of the medium alone does not influence basolateral APP secretion (B). The apical secretion in the absence of methylamine was always lower than 5% and is not displayed. Every point is the mean ± S.E. of three separate experiments in which the secretion in every compartment was compared with the total secretion of APP under control conditions.

The effect of methylamine was concentration-dependent (Fig. 4, a and b). Total metabolic incorporation of radiolabeled methionine in trichloroacetic acid-precipitable protein in the presence of 10 or 30 mM methylamine was 100.8 and 99.2%, respectively, of the control, indicating that cell viability and metabolism, as measured by overall protein synthesis, was not affected. Finally, the effect of methylamine was reversible after 1 h of incubation with normal medium (Fig. 4c).

Figure 4: Effect of primary amines, bafilomycin A1, and other drugs on polarized APP secretion. Panel a, MDCK cells transfected with APP695(M) were preincubated (30 min) and metabolically labeled (120 min) in medium containing the following drugs: no additions (control), 0.2 mM monodansylcadaverine (MDCV), 20 mM 6-amino-1-hexanol (Am.Hex), 40 mM glycylglycine (Glygly), 30 mM methylamine (30 mM MA), 10 mM methylamine (10 mM MA), 30 mM ammonium chloride (30 mM NH), 10 mM ammonium chloride (10 mM NH), 0.3 mM chloroquine (chlq). Double immune precipitation was performed with antibody B2/3. Panel b, Untransfected MDCK cells were treated as in panel a with no drug (control), 10 mM NHCl, 10 or 30 mM methylamine (MA), or 100 mM and 1 µM bafilomycin A1 (Baf A1). Bars represent the mean (with S.E.) apical (A) and basolateral (B) secretion of canine APP in three separate experiments. The total secretion (apical + basolateral) is taken as 100%. Panel c (upper part), untransfected MDCK cells were preincubated and labeled in the presence of 20 mM methylamine (MA), 100 nM bafilomycin A1 (Baf), or 40 mM glycyl-glycine (Glygly). Cont, control, with no drugs added. Apical (A) and basolateral (B) medium was recovered and immune precipitated with antibody 207. Panel c (lower panel), cells were further incubated with medium containing no drugs during 1 h. A second round of labeling (30 min of preincubation, 2 h of labeling) was performed without any addition of drugs. Immunoprecipitation of the medium was again performed with antibody 207. Note the presence of a 65-kDa protein in the apical and basolateral compartment appearing after washing out bafilomycin A1. This represents a proteolytic fragment of APP, the significance of which is unclear at this moment.

Other primary amines such as ammonium chloride (10 or 30 mM), ammonium acetate (20 mM), and 6-amino-1-hexanol (10 or 30 mM) caused secretion of endogenous APP751/770 and of transfected APP695 in the apical compartment (Fig. 4a). Glycyl-glycine (20-40 mM), chloroquine (0.3 mM), and monodansylcadaverine (0.1 or 0.2 mM), in contrast, did not significantly affect the polarized secretion of APP toward the basolateral compartment of MDCK cells (Fig. 4a).

Other drugs such as forskolin (10 µM), cholera toxin (2 µg/ml), and pertussis toxin (2 µg/ml) did not significantly affect the polarized secretion of APP in the conditions used, whereas the phorbol esters PMA (1 µM) and Pdbu (1 µM) were confirmed to increase the basolateral secretion of APP (34) (results not shown).

The known alkalization by primary amines of intracellular compartments, such as the basolateral endosome, the lysosomes, or the trans-Golgi network, can be mimicked by drugs of the macrolide family that inhibit specifically the vacuolar proton ATPase(44) . Incubation of MDCK cells with bafilomycin A1 (10 nM, 100 nM, and 1 µM), bafilomycin B1 (10 nM and 100 nM), concanamycin A (10 nM, 100 nM, and 1 µM), and concanamycin C (100 nM) essentially randomized the secretion of APP to the apical and the basolateral compartment (Fig. 4b, and results not shown). The effect of bafilomycin A1 was not reversed by incubation for 1 h without the drug (Fig. 4c). A possible explanation is the high affinity of bafilomycin A1 for the vacuolar H-ATPase, although this was not further investigated. The concentrations used here are in the same range as previously demonstrated by others to inhibit specifically the vacuolar proton pump(44, 45) .

Deletion of the Cytoplasmic and Transmembrane Domains of APP

Figure 5: Missorting of deletion mutants of APP695(M) in MDCK cells. Panels a and b, MDCK cells were transfected with wild type APP695(M) (wild type), APP695(M)A666* (A666*), APP695(M)Y653* (Y653*), APP695(M)A626* (A626*), APP695(M)L613* (L613*), and APP695(M)D597* (D597*), respectively; see Fig. 1for a schematic representation of these mutants. Cells were labeled for 2 h, and immune precipitation was performed with antibody 207. Apical (A) secretion of APP (10-18%) was observed after transfection with mutant APP terminating at the -secretase cleavage site (D597*), at the -secretase cleavage site (L613*), or at the transmembrane domain (A626*). Panel c, a high expression MDCK clone transfected with APP695(M)A666* was obtained by limiting dilution. Cells were labeled for 2 h and APP in the apical and basolateral medium was immune precipitated with antibody B2/3. MDCK cells overexpressing wild type APP695 and transfected with pRC/RSV plasmid (control), containing no insert (see Fig. 2b) are also shown. Whereas the expression level of the transfected constructs in these cells is at least 10-fold higher than that of endogenous APP, both transfected wild type APP695(M) and APP695(M)A666* are secreted exclusively into the basolateral compartment.

To investigate the role of the amyloid domain in the polarized secretion of APP, stop codons were introduced at positions Leu (APP695(M)L613*) and Asp (APP695(M)D597*), to produce soluble APP terminating at the - and at the -secretase cleavage sites, respectively(6) . Both soluble APP forms were still observed to be mainly secreted into the basolateral compartment, although 10-15% of the total secreted APP was sorted toward the apical compartment (Fig. 5b).

Sorting of APP Containing the Double Swedish Mutation

Figure 6: Apical missorting after transfection with the swedish EOFAD APP mutant. Panel a, MDCK cells were transfected with APP695(H)617A:G (617A:G), APP695(H)618E:Q (618E:Q), APP695(H)595K:N/596 M:L (Swed, swedish mutation), and APP695(H)642V:I (642V:I). These APP mutations cause familial Alzheimer's Disease or hereditary cerebral hemorrhage with amyloidosis-Dutch type. Labeling was performed overnight, and double immune precipitations on apical (A) and basolateral (B) compartments were performed with antibody 207. Apical secretion (10-20%) is observed only by transfection with the swedish mutant APP. Panel b, apically secreted APP containing the swedish mutation is generated by -secretase. MDCK cells transfected with APP695(H)595K:N/596 M:L (swedish mutation) were labeled for 2 h. Apical (A) and basolateral (B) medium were consecutively immune precipitated two times with antibody R1736, which recognizes residues 2-15 of the -amyloid peptide and therefore precipitates only -secretase-cleaved APP. Antibody 207, which recognizes both - and -secretase-cleaved APP, was used in one further round of immune precipitation to recover -secretase-cleaved APP. Note that APP containing the swedish mutation and secreted into the apical compartment is only precipitated with antibody 207. Most -cleaved APP, however, remains secreted into the basolateral compartment. Note the slight difference in mobility of -secretase-cleaved APP in the apical compartment in panel a.

In other transfection experiments, MDCK cells were obtained in which no apical missorting of the swedish APP-mutant was observed (result not shown). Further analysis demonstrated that in these cases also no -secretase activity (Fig. 6) could be demonstrated. Although further investigations are needed to explain this variation in -secretase activity, the conclusion that only -secretase-cleaved APP is missorted is supported by this observation.

DISCUSSION

We investigated the secretion of APP in polarized MDCK cells. This cell line is by far the best studied model system for protein sorting (29, 30) . Moreover, certain aspects of the sorting mechanism operate similarly to that of hippocampal neurons(31, 32) . The physical separation between apical and basolateral compartments in MDCK cells cultured on polycarbonate filters permits the analysis and comparison of sorting of wild type and mutant APP as well as the study of drugs in a qualitative and quantitative fashion. This is currently not feasible in primary cultures of hippocampal neurons because of practical limitations.

The major observations are, first, that endogenously produced canine APP is secreted in a strictly polarized fashion in MDCK cells(34) , being sorted to the basolateral compartment by a mechanism that is characterized by its high fidelity and capacity. More than 10-fold overexpression of wild type APP695 did not result in any apparent apical secretion. Advantage of this observation can be taken by using the basolateral secretion of endogenous canine APP751/770 as an internal control and marker for the basolateral compartment. This result also demonstrates that APP (which is transported by fast axonal transport in neurons) is an exception to the rule that basolateral sorted proteins in MDCK cells are sorted somatodendritically in neurons (31, 32, 33) . Second, we demonstrate the extreme sensitivity of the sorting mechanism to intracellular alkalization. This indicates that the sorting mechanism resides in an intracellular acid compartment, e.g. the trans-Golgi network or the basolateral endosome, and that sorting per se is a pH-sensitive process. Third, deletion of the cytoplasmic domain with the NPTY endocytosis and potential basolateral sorting signal (46, 47) did not markedly affect the basolateral secretion of APP. Deletion of nearly the entire cytoplasmic domain by introducing a stop codon at Tyr affected the polarized secretion of APP also only marginally (5-14%). Although the putative sorting signals in the cytoplasmic domain could be important for the targeting of uncleaved, membrane-bound APP, our results demonstrate that the fraction of APP destined for secretion is sorted by means of a determinant in the extracellular domain. This is confirmed by the results obtained with soluble APP mutants terminating at the transmembrane domain, at the -cleavage site, or at the -cleavage site. Fourth, all of the soluble mutants tested displayed a small but significant leakage to the apical compartment, amounting to 10-18% of the total pool of secreted APP. The direct consequence is that about 5-20-fold more soluble APP is accumulating in the apical compartment, compared with wild type APP. Finally, we observed that a fraction of soluble, -secretase-cleaved APP was targeted to the apical compartment after introduction of the swedish point mutations (12) in the APP protein.

In MDCK cells, integral membrane and soluble proteins destined for the apical or the basolateral surface are sorted in the trans-Golgi network (TGN)(28, 29, 30) . A second sorting site is the basolateral endosome, from which endocytosed basolateral proteins can recycle to the basolateral domain, travel to the lysosomes, or transcytose to the apical domain (29, 30) . Similar basolateral sorting signals are operative in the TGN and the basolateral endosome(47) . Since APP is synthesized as an integral membrane protein to become processed into a soluble form, it is unclear whether APP is sorted as an integral membrane protein (in which case the cytoplasmic domain is expected to contain the major basolateral sorting determinants) or as a soluble protein (in which case the sorting determinants must reside in the extracellular domain). The subcellular localization of the processing -secretase is crucial in this regard: if cleavage of APP occurs before the TGN is reached, APP would be sorted as a soluble protein. Our results obtained with the cytoplasmic deletion mutants and with the soluble APP mutants terminating at the - and the -secretase cleavage site provide clear evidence for a major basolateral sorting determinant in the extracellular domain of APP: all APP mutants, soluble or not, remain essentially targeted to the basolateral compartment. Moreover, addition of the broad spectrum proteinase inhibitor 2-macroglobulin (48) in both apical and basolateral compartments did not affect APP secretion, which is in agreement with the idea that the bulk of -secretase activity in MDCK cells is located intracellularly(21, 22, 23, 24) .

The soluble mutants (APP695(M)A626*, APP695(M)L613*, and APP695(M)D597*) are all, however, less efficiently sorted than wild type APP. The apical sorting of a fraction (10-18%) of soluble APP is not the consequence of overexpression of the mutants, since low expression of the APP695(M)L613 mutant yielded similar results, and a more than 10-fold overexpression of wild type APP695 and APP695(M)Ala666* did not cause apical secretion. Moreover, the measured impermeability of the cell layer for [H]inuline and the exclusive basolateral secretion of endogenous canine APP751/770 excluded leakage from the basolateral to the apical compartment in these experiments. As stated above, this endogenous APP functions as a nearly perfect marker for the basolateral secretion compartment of MDCKII cells. Some apical missorting of soluble APP via apical transport vesicles in the TGN, following the bulk of soluble proteins that are secreted apically in MDCK cells, is therefore the most likely explanation.

Methylamine and other primary amines disturb secretion and sorting of APP in MDCK cells, resulting in the preferential secretion of APP into the apical compartment. At the concentrations used, the metabolic incorporation of, and labeling by [S]methionine was not affected, and the cell layer impermeability was not significantly altered. The drugs were easily washed out and the basolateral secretion restored by incubating the cells in normal medium for a relatively short time period. The same effect is observed with related drugs such as ammonium chloride, ammonium acetate, and 6-amino-1-hexanol. Glycyl-glycine, used as a control, did not affect the polarity of APP secretion. The effect of methylamine was accentuated by alkalization of the medium, which also caused an unexpected substantial increase in the total amount of (mainly apically) secreted APP. This increase in secretion is most easily explained by blocking an as yet undefined but degradation-oriented pathway and redirection of APP into the apical secretory pathway. This is corroborated by data showing that treatment of MDCK cells with primary amines results in randomized sorting of lysosomal proteinases and secretion in the extracellular medium(43) . In unpolarized cells an unknown fraction of APP is targeted to the lysosomes(2) . Whereas the inhibition of basolateral APP secretion in MDCK cells after methylamine treatment is in line with the inhibition of APP secretion observed in unpolarized cells(21) , the stimulation of APP secretion along the apical pathway in MDCK cells suggests differences between the apical and the basolateral secretory transport machinery, at least as far as APP is concerned. It is clear that the effect of methylamine in MDCK cells is partially mimicked by bafilomycins and concanamycins, drugs that specifically block the vacuolar H-ATPase(44) . Therefore, the combined observations with primary amines and specific H-ATPase inhibitors in MDCK cells demonstrate that an acid compartment is involved in the correct sorting of APP and that at least part of the missorting of APP is explained by alkalization of this compartment. It is likely that binding of the ectodomain of APP to receptors or chaperones that are responsible for its specific basolateral sorting are pH-sensitive. This phenomenon is essential for recycling of endocytosis receptors (50) and was also postulated for the basolateral sorting of soluble laminin and proteoglycan in MDCK cells(43) . On the other hand, recent evidence suggests that maintenance of the acid pH of microsomal vesicles is essential to bind small GTP binding proteins and ADP ribosylation factor, which are involved in the specific coupling of transport vesicles to their target membranes(51, 52) . The possibility that vesicular alkalization interferes with such incompletely understood targeting mechanisms is therefore an alternative interpretation, which can be approached experimentally.

Interestingly, -secretase cleavage of the swedish mutant APP, as opposed to -secretase cleavage of normal APP, resulted in apical leakage of a fraction of soluble APP. The results obtained with soluble APP mutants terminating at either the -secretase cleavage site or the -secretase cleavage site, rule out the possibility that the basolateral sorting signal is in the A4 amyloid sequence itself. The missorting of truncated soluble APP mutants, and of -cleaved APP with the swedish mutation suggests, therefore, that merely the soluble character (defined as absence of the cytoplasmic and integral membrane domain) is responsible for the partial mistargeting of APP to the apical compartment, in line with what we discussed above. The fact that the -secretase-cleaved swedish mutant APP was mistargeted to the apical compartment (10-20%) to a similar degree as the synthetic soluble APP mutants is compatible with the hypothesis that the swedish mutant is cleaved by -secretase early in the biosynthetic pathway, before the TGN is reached. Its further behavior is then identical to that of the soluble transfectants.

In conclusion, our results demonstrate that small structural or metabolic persistent disturbances of the protein-sorting machinery of polarized cells could be involved in the abnormal processing of amyloid precursor protein. The possibility that missorting of APP plays a role in the pathogenesis of certain forms of Alzheimer's Disease requires further investigation.

FOOTNOTES

*

This investigation was supported by Grants 3.0069.89 and 3.0073.93 from the Fonds voor Geneeskundig Wetenschappelijk Onderzoek, by EC Contract BIOT-CT91-0302, by STW Contract NCH22.2726, by a grant Geconcerteerde Acties from the Ministerie voor Onderwijs of the Belgian Government, and by a grant from the interuniversity network for Fundamental Research (IUAP, 1991-1996). Part of this work was done under contract with the Action Program for Biotechnology of the Flemish government (VLAB-IWT, ETC-008). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore by hereby marked ``advertisement'' in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

§

Postdoctoral fellow of the Nationaal Fonds voor Wetenschappelijk Onderzoek.

¶

To whom correspondence should be addressed: Experimental Genetics Group, Dept. of Human Genetics, Campus Gasthuisberg O & N 6, B-3000 Leuven, Belgium. Tel.: 32-16-3458-62; Fax: 32-16-3458-71; fredvl{at}cc3.kuleuven.ac.be.

()

The abbreviations used are: APP, amyloid precursor protein; AD, Alzheimer's disease; EOFAD, early onset familial Alzheimer's disease; TGN, trans-Golgi network; MDCK, Madin-Darby canine kidney; PMA, phorbol 12-myristate 13-acetate; Pdbu, phorbol 12,13-dibutyrate; TES, 2-{[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]amino}ethanesulfonic acid; RSV, Rous sarcoma virus.

ACKNOWLEDGEMENTS

Addendum-After this paper was submitted, we learned that the work of C. Haass et al.(53) confirms our findings completely.

REFERENCES

1. Weidemann, A., König, G., Bunke, D., Fischer, P., Salbaum, J. M., Masters, C. L., and Beyreuther, K. (1989) Cell 57, 115-126 [CrossRef][Medline] [Order article via Infotrieve]
2. Estus, S., Golde, T., Kunishita, T., Blades, D., Lowery, D., Eisen, M., Usiak, M., Qu, X., Tabira, T., Greenberg, B., and Younkin, S. (1992) Science 255, 726-728 [Abstract/Free Full Text]
3. Golde, T., Estus, S., Younkin, S., Selkoe, D., and Younkin, S. (1992) Science 255, 728-730 [Abstract/Free Full Text]
4. Haass, C., Schlossmacher, M. G., Hung, A., Vigo-Pelfrey, C., Mellon, A., Ostaszewski, B., Lieberburg, I., Koo, E., Schenk, D., Teplow, D., and Selkoe, D. (1992) Nature 359, 322-325 [CrossRef][Medline] [Order article via Infotrieve]
5. Shoji, M., Golde, T., Ghiso, J., Cheung, T., Estus, S., Shaffer, L., Cai, X. D., McKay, D., Tintner, R., Frangione, B., and Younkin, S. (1992) Science 258, 126-129 [Abstract/Free Full Text]
6. Haass, C., and Selkoe, D. (1993) Cell 75, 1039-1042 [CrossRef][Medline] [Order article via Infotrieve]
7. Glenner, G., and Wong, C. (1984) Biochem. Biophys. Res. Comm. 120, 885-890 [CrossRef][Medline] [Order article via Infotrieve]
8. Masters, C., Simms, G., Weinman, N., Multhaup, G., McDonald, B., and Beyreuther, K. (1985) Proc. Natl. Acad. Sci. U. S. A. 82, 4245-4249 [Abstract/Free Full Text]
9. Levy, E., Carman, I. J., Fernandez-Madrid, M. D., Power, I., Lieberburg, I., Van Duynen, G. T., Bots, A. M., Luyendijk, W., and Frangione, B. (1990) Science 248, 1124-1126 [Abstract/Free Full Text]
10. Van Broeckhoven, C., Haan, J., Bakker, E., Hardy, J., Van Hul, W., Wehnert, A., Vegter-Van Der Vlis, M., and Roos, R., (1990) Science 248, 1120-1122 [Abstract/Free Full Text]
11. Goate, A., Chartier-Harlin, M. C., Mullan, M., and 18 coauthors (1991) Nature 349, 704-706 [CrossRef][Medline] [Order article via Infotrieve]
12. Mullan, M., Crawford, F., Axelman, K., Houlden, H., Lilius, L., Winblad, B., and Lannfelt, L. (1992) Nat. Genet. 1, 345-347 [CrossRef][Medline] [Order article via Infotrieve]
13. Hendriks, L., van Duijn, C. M., Cras, P., Cruts, M., Van Hul, W., van Harskamp, F., Warren, A., McInnis, M. G., Antonorakis, S. E., Martin, J. J., Hofman, A., and Van Broeckhoven, C. (1992) Nat. Genet. 1, 218-222 [CrossRef][Medline] [Order article via Infotrieve]
14. Rumble, B., Retallack, R., Hilbich, C., Simms, G., Multhaup, G., Martins, R., Hockey, A., Montgomery, P., Beyreuther, K., and Masters, C. L. (1989) New Engl. J. Med. 320, 1446-1452 [Abstract]
15. Yankner, B., Caceres, A., and Duffy, L. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 9020-9023 [Abstract/Free Full Text]
16. Pike, C., Burdick, J., Walencewicz, D., Glabe, A. J., and Cotman, C. W. (1993) J. Neurosci. 13, 1676-1687 [Abstract]
17. Esch, F. S., Keim, P. S., Beattle, E. C., Blacher, R. W., Culwell, A. R., Oltersdorf, T., McClure, D., and Ward, P. J. (1990) Science 248, 1122-1124 [Abstract/Free Full Text]
18. Zhong, Z., Higaki, J. H., Murakami, K., Wang, Y., Catalano, R., Quon, D., and Cordell, B. (1994) J. Biol. Chem. 269, 627-632 [Abstract/Free Full Text]
19. Maruyama, K., Kametani, F., Usami, M., Yamao-Harigaya, W., and Tanaka, K. (1991) Biochem. Biophys. Res. Commun. 179, 1670-1676 [CrossRef][Medline] [Order article via Infotrieve]
20. Sisodia, S. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 6075-6079 [Abstract/Free Full Text]
21. De Strooper, B., Umans, L., Van Leuven, F., and Van den Berghe, H. (1993) J. Cell Biol. 121, 295-304 [Abstract/Free Full Text]
22. De Strooper, B., Van Leuven, F., and Van den Berghe, H. (1992) FEBS Lett. 308, 50-53 [CrossRef][Medline] [Order article via Infotrieve]
23. Sambamurti, K., Shioi, J., Anderson, J. P., Pappolla, M. A., and Robakis, N. (1992) J. Neurosci. Res. 33, 319-329 [CrossRef][Medline] [Order article via Infotrieve]
24. Kuentzel, S., Ali, S., Altman, R., Greenberg, B., and Raub, T. (1993) Biochem. J. 295, 367-378
25. Martin, G. M., Schellenberg, G. D., Wijsman, E. M., and Bird, T. D. (1990) Nature 347, 124 [CrossRef][Medline] [Order article via Infotrieve]
26. Seiler, N. (1993) Neurochem. Res. 18, 235-245 [CrossRef][Medline] [Order article via Infotrieve]
27. Clark, R. F., and Goate, A. (1993) Arch. Neurol. 50, 1164-1172 [CrossRef][Medline] [Order article via Infotrieve]
28. Simons, K., and Wandinger-Ness, A. (1990) Cell 62, 207-211 [CrossRef][Medline] [Order article via Infotrieve]
29. Rodriguez-Boulan, E., and Powell, S. (1992) Annu. Rev. Cell Biol. 8, 395-427 [CrossRef]
30. Simons, K., and Zerial, M. (1993) Neuron 11, 789-799 [CrossRef][Medline] [Order article via Infotrieve]
31. Dotti, C., and Simons, K. (1990) Cell 62, 63-72 [CrossRef][Medline] [Order article via Infotrieve]
32. Dotti, C., Parton, R., and Simons, K. (1991) Nature 349, 158-161 [CrossRef][Medline] [Order article via Infotrieve]
33. Huber, L., De Hoop, M., Dupree, P., Zerial, M., Simons, K., and Dotti, C. (1993) J. Cell Biol. 123, 47-55 [Abstract/Free Full Text]
34. Haass, C., Koo, E., Teplow, D., and Selkoe, D. J. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 1564-1568 [Abstract/Free Full Text]
35. Citron, M., Oltersdorf, T., Haass, C., McConlogue, L., Hung, A. Y., Seubert, P., Vigo-Pelfrey, C., Lieberburg, I., and Selkoe, D. J. (1992) Nature 360, 672-674 [CrossRef][Medline] [Order article via Infotrieve]
36. Cai, X. D., Golde, T., and Younkin, S. G. (1993) Science 259, 514-516 [Abstract/Free Full Text]
37. Gillespie, S. L., Golde, T. E., and Younkin, S. G. (1992) Biochem. Biophys. Res. Commun. 187, 1285-1290 [CrossRef][Medline] [Order article via Infotrieve]
38. De Strooper, B., Van Leuven, F., and Van den Berghe, H. (1991) Biochim. Biophys. Acta 1129, 141-143 [Medline] [Order article via Infotrieve]
39. Urban, J., Parczyk, K., Leutz, A., Kayne, M., and Kondor-Koch, C. (1987) J. Cell Biol. 105, 2735-2743 [Abstract/Free Full Text]
40. Deng, W. P., and Nickoloff, J. A. (1992) Anal. Biochem. 200, 81-88 [CrossRef][Medline] [Order article via Infotrieve]
41. De Strooper, B., Van Leuven, F., Carmeliet, G., Van den Berghe, H., and Cassiman J. J. (1991) Eur. J. Biochem. 199, 25-33 [Medline] [Order article via Infotrieve]
42. Dyrks, T., Dyrks, E., Mönning, U., Urmoneit, B., Turner, J., and Beyreuther, K. (1993) FEBS Lett. 335, 89-93 [CrossRef][Medline] [Order article via Infotrieve]
43. Caplan, M., Stow, J., Newman, A., Madri, J., Anderson, H. C., Farquhar, M., Palade, G., and Jamieson, J. D. (1987) Nature 329, 632-635 [CrossRef][Medline] [Order article via Infotrieve]
44. Bowman, E. J., Siebers, A., and Altendorf K. (1988) Proc. Natl. Acad. Sci. U. S. A. 85, 7972-7976 [Abstract/Free Full Text]
45. Yoshimori, T., Yamamoto, A., Moriyama, Y., Futai, M., and Tashiro, Y. (1991) J. Biol. Chem. 266, 17707-17712 [Abstract/Free Full Text]
46. Matter, K., Hunziker, W., and Mellman, I. (1992) Cell 71, 741-753 [CrossRef][Medline] [Order article via Infotrieve]
47. Matter, K., Whitney, J. A., Yamamoto, E. M., and Mellman, I. (1993) Cell 74, 1053-1064 [CrossRef][Medline] [Order article via Infotrieve]
48. Van Leuven, F. (1982) Trends Biochem. Sci. 7, 185-187
49. Deleted in proof
50. Van Leuven, F., Cassiman, J. J., and Van den Berghe, H. (1980) Cell 20, 37-43 [CrossRef][Medline] [Order article via Infotrieve]
51. Zeuzem, S., Feick, P., Zimmermann, P., Haase, W., Kahn, R. A., and Schulz, I. (1992) Proc. Natl. Acad. Sci. U. S. A. 89, 6619-6623 [Abstract/Free Full Text]
52. Zeuzem, S., Zimmermann, P., and Schulz, I. (1992) J. Membr. Biol. 125, 231-241 [Medline] [Order article via Infotrieve]
53. Haass, C., Koo, E. H., Capell, A., Teplow, D. B., and Selkoe, D. J. (1995) J. Cell Biol. , in press

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				A. Uelhoff, J. Tatzelt, A. Aguzzi, K. F. Winklhofer, and C. Haass A Pathogenic PrP Mutation and Doppel Interfere with Polarized Sorting of the Prion Protein J. Biol. Chem., February 18, 2005; 280(7): 5137 - 5140. [Abstract] [Full Text] [PDF]

				A. Capell, L. Meyn, R. Fluhrer, D. B. Teplow, J. Walter, and C. Haass Apical Sorting of beta -Secretase Limits Amyloid beta -Peptide Production J. Biol. Chem., February 8, 2002; 277(7): 5637 - 5643. [Abstract] [Full Text] [PDF]

				J. Wiltfang, H. Esselmann, P. Cupers, M. Neumann, H. Kretzschmar, M. Beyermann, D. Schleuder, H. Jahn, E. Ruther, J. Kornhuber, et al. Elevation of beta -Amyloid Peptide 2-42 in Sporadic and Familial Alzheimer's Disease and Its Generation in PS1 Knockout Cells J. Biol. Chem., November 9, 2001; 276(46): 42645 - 42657. [Abstract] [Full Text] [PDF]