Abstract
The secretory N-terminal domain of the Alzheimer amyloid precursor protein (sAPP) evokes specific responses in cells on binding to their surfaces. Because APP is expressed in a large variety of cell types, the localization of sAPP binding requires detection techniques that selectively recognize sAPP as a ligand. For this purpose, we prepared antibodies against recombinant sAPP695 (sAPPrec) previously expressed in E. coli. Such antibodies were found to distinguish between sAPPrec and cellular APP or sAPP, as shown by immunocytochemistry and by immunoblot. In addition, they allowed the selective localization of bound sAPPrec on cell surfaces without any signal from cellular APP or sAPP. Saturation of sAPPrec binding to cell surfaces, as determined radiometrically, was reached at 10 nM [125 I]-sAPPrec. Binding was specific because it was almost completely inhibited by a 100-fold excess of unlabeled sAPPrec. This specificity of binding was confirmed by surface plasmon resonance spectroscopy. Binding of sAPPrec to cell surfaces occurred in patches and was dependent on the state of cell differentiation. The sAPPrec used in this study contains heparin binding sites, but enzymatic removal of cell surface associated heparin did not affect sAPPrec binding. Aldehyde fixation of cells strongly inhibited their ability to bind sAPPrec. The data point to a fixation-sensitive sAPPrec binding protein which is detectable in the form of patches and therefore is part of assembled cell surface microdomains.
Keywords
T
The existence of a membrane receptor for sAPP has been postulated from the detection of sAPP signal transduction mechanisms observed in a variety of cell types (Greenberg et al. 1994; Jin et al. 1994; Barger and Mattson 1996; Pietrzik et al. 1998). This putative sAPP receptor is of great interest but is entirely unknown at present. The LDL receptor-related protein has been shown to bind sAPP770 through its Kunitz type protease inhibitor (KPI) domain (Kounnas et al. 1995). However, this receptor protein cannot be considered as a general sAPP receptor because the KPI domain is lacking in a number of sAPP species (Golde et al. 1990; Kang and Müller-Hill 1990).
Direct light or electron microscopic cellular localization of receptor molecules is usually performed by the use of antibodies directed against this receptor (for review see Polak 1988; Sandor et al. 1994; Elde et al. 1995). Because the receptor for sAPP is as yet unidentified, indirect techniques for the localization of binding sites, e.g., by the use of antibodies against sAPP, are necessary. However, the ubiquitous expression of cellular APP interferes with the immunocytochemical localization of exogenously added sAPP bound to cell surfaces. To solve this problem, techniques for selective visualization of cell surface-bound sAPP are required to indirectly localize cellular sAPP binding sites.
The aim of this study was to develop suitable techniques for the cytochemical detection of cell surface-bound sAPP. For this purpose, recombinant sAPP (sAPPrec) containing a His tag was expressed in E. coli, purified, and biotinylated, or was used as an antigen. To exclude that cell surface binding of sAPPrec might be mediated by the KPI sequence (see above), we used sAPPrec derived from APP 695 lacking this domain. The detection techniques, which involved either streptavidin or antibodies against sAPPrec or against its His tag, were able to efficiently discriminate between sAPPrec and cellular APP. Our results indicate that anti-sAPPrec antibodies provided the most sensitive detection technique for cell surface-bound sAPPrec compared to both of the other techniques. We show that binding of sAPPrec occurs in the form of microdomains and that it is saturable and specific.
Materials and Methods
Materials and Cells
FRTL-5 (Ambesi-Impiombato et al. 1980) and PC 12 (Greene and Tischler 1976) cells were obtained from American Type Culture Collection (ATCC; Rockville, MD). B104 cells (Schubert et al. 1974) were kindly provided by Dr. G. Schwarzmann (University of Bonn, Germany). Polyclonal antiserum 3329 (Pietrzik et al. 1998) against the recombinant His6-sAPP695 (sAPPrec) and antiserum 2189 (Popp et al. 1996) against the APP C-terminus were raised in rabbits as described. Anti-sAPPrec antibodies from chicken were used for double labeling experiments. The chicken antibodies were raised according to Gassmann et al. (1990) and purified as described by Lösch et al. (1986). Preimmune serum was taken from each animal and was used later for immunocytochemical control experiments. The mouse monoclonal antibody (MAb) 22C11 against the denatured N-terminus of APP (Weidemann et al. 1989) was provided by K. Beyreuther and G. Multhaup (University of Heidelberg, Germany). The anti-His4 antibody was obtained from Quiagen (Hilden, Germany). NHS-biotin and heparinase I were obtained from Sigma (Deisenhofen, Germany). Expression and purification of sAPPrec were performed as described previously (Pietrzik et al. 1998). Collagen was isolated from rat tail tendon, dried, sterilized by UV radiation, and solubilized in 0.1% acetic acid (Strom and Michalopoulos 1982). Collagen-coated coverslips were prepared as previously described (Popp et al. 1996).
Cell Culture
FRTL-5 cells were cultured at 37C with 5% CO2/95% air in F-12 medium (Coons' modification; Seromed, Berlin, Germany) supplemented with 5% calf serum, TSH (1 milliU/ml), insulin (10 μg/ml), hydrocortisone (10 μM), Gly-His-Lys (10 ng/ml), somatostatin (10 ng/ml), transferrin (5 μg/ml), penicillin (50 U/ml), and streptomycin (50 U/ml). B104 or PC 12 cells were cultured at 37C and 5% CO2/95% air in Dulbecco's minimal essential medium (DMEM; Gibco, Eggenstein, Germany) supplemented with 10% fetal calf serum (FCS). Differentiation of B104 cells was induced by culturing cells for 3 days on collagen-coated coverslips in serumfree DMEM containing 1 mM db-cAMP (Sigma).
Light Microscopic Detection of Endogenous APP in FRTL-5, B104, and PC12 Cells
To localize APP, FRTL-5, B104, or PC12 cells were seeded on coverslips coated with collagen. Cells were fixed in paraformaldehyde, permeabilized with 0.2% Triton X-100, blocked for 30 min with 3% BSA in PBS, and labeled with a rabbit serum against the C-terminal region of APP diluted in PBS with 0.3% BSA (PBSA). For immunofluorescence detection, the cells were incubated with DTAF-labeled goat anti-rabbit IgG (Dianova; Hamburg, Germany) diluted 1:50 in PBSA for 60 min at 37C and were viewed with a TCS 4D Leica confocal laser scanning microscope (Leica; Bensheim, Germany).
Binding of Exogenously Added sAPPrec to the Surface of FRTL-5, B104, or PC12 Cells
To analyze the binding properties of sAPPrec, FRTL-5 or B104 cells were seeded on coverslips and incubated with sAPPrec. For specific detection of bound sAPPrec the rabbit anti-sAPPrec antibody (antiserum 3329) exclusively recognizing sAPPrec but not endogenous APP was used. Cells were blocked for 30 min with 3% BSA in PBS for 60 min at 4C and then incubated with 0.06, 0.1, 1, 10, or 100 nM sAPPrec, which was detected by immunofluorescence as described above. The concentration of 60 pM sAPPrec was included in our experiments because this concentration might be relevant for cells exposed to the bloodstream, in which this sAPP level has been reported to appear (Van Nostrand et al. 1991). Double labeling experiments for simultaneous visualization of endogenous APP and exogenously added sAPPrec were performed by the use of rabbit antiserum 2189 directed against the APP C-terminus (see above) and the chicken anti-sAPPrec antibody B1. The chicken anti-sAPPrec antibody was detected by a DTAF-labeled monoclonal mouse anti-chicken antibody (Sigma), whereas for the antibody 2189 a TRITC-labeled goat anti-rabbit antibody (Dianova) was used. For the heparinase assay, cells on cover-slips were treated with different concentrations of heparinase I in heparinase buffer for 2 hr at 37C before the binding assay was performed. The heparinase buffer contained Tris (20 mM), NaCl (50 mM), CaCl2 (4 mM), and BSA (0.01%). Densitometric measurements of fluorescence intensities were performed using standard software of the confocal microscope (Leica). The results were compared with the staining obtained by the use of the anti-His tag antibody (anti-His4; Quiagen, Hilden, Germany) and with biotinylated sAPPrec as recognized by the use of streptavidin-Cy3 (Sigma). sAPP was biotinylated using the technique described by Bayer et al. (1979).
Control Experiments
Several control experiments for each labeling procedure were performed to assess the immunocytochemical specificity. Controls included the substitution of antiserum by rabbit or chicken preimmune serum to evaluate the specificity of the primary antibody. To evaluate nonspecific binding of the secondary fluorescence labeled antibodies, control experiments with omission of the primary antibody were performed. The inhibitory effect of free antigen was tested by adsorption of the specific antibodies to sAPPrec before immunolabeling. For this purpose, diluted rabbit or chicken antisera were preincubated overnight with an excess of purified sAPPrec and were then used for immunocytochemical detection.
Transmission Electron Microscopy
For transmission electron microscopy, cells were seeded on plastic coverslips and incubated with sAPPrec, followed by the primary antibody as described above. As secondary antibody, a gold-labeled goat anti-rabbit IgG (Dianova) was used, and the gold label (5 nm) was enhanced using a silver enhancement kit (IntenSE; Amersham, Freiburg, Germany). Cells were postfixed for 10 min with 1% osmium tetroxide, stained en bloc with 2% aqueous uranyl acetate, embedded in Epon 820 (Fluka; Deisenhofen, Germany), and mounted on plastic coverslips. Thin sections were stained with lead citrate (10 min) and examined with a Philips CM120 electron microscope (Philips Electron Optics; Eindhoven, Netherlands).
SDS-PAGE and Immunoblotting
FRTL-5 cells were cultured for 48 hr in serum-free F-12 medium. Medium was collected at 4C and concentrated with a SpeedVac evaporation device (SC100; Savant, Farmingdale, NY). Tenfold concentrated medium was boiled for 5 min in sample buffer and proteins were separated on a 12.5% reducing SDS gel (Schägger and von Jagow 1987) and transferred to nitrocellulose. The blots were immersed overnight in blocking solution (6% casein in 1% polyvinylpyrrolidone 40 with 10 mM EDTA in PBS, pH 6.8) and incubated with the mouse anti-APP MAb 22C11 or the antiserum 3329, followed by incubation with the peroxidase-labeled secondary anti-mouse or anti-rabbit IgG antibody (Dianova). APP from cell lysates (10 μg per lane) or recombinant sAPP (5 μg per lane) was analyzed by use of the mouse anti-APP MAb 22C11 or the rabbit anti-sAPPrec antibody 3329. The APP-specific bands were visualized by chemiluminescence (ECL; Amersham) and documented on XAR-5 films (Kodak; Stuttgart, Germany).
Determination of sAPPrec Interaction with Cells
Radiometric Quantitation
For quantitation of sAPPrec binding to cells, sAPPrec was iodinated with [125 I]-NaI using iodobeads (Pierce; Oud-Beijer, The Netherlands) for 15 min at RT (Markwell 1982). Iodinated sAPPrec was desalted (Econo Pac 10 DG; Bio-Rad Laboratories, Hercules, CA). This iodination procedure yielded a specific radioactivity of 400 cpm/ng protein. For the binding assay, FRTL-5 cells were grown in 6-well plates in F-12 medium (see above). Cells were washed at 4C with ligand binding buffer (PBS with 200 μM CaCl2 and 10 mM ∊-amino capronic acid) and incubated with ligand binding buffer containing 0.1 to 100 nM [125 I]-sAPPrec for 90 min at 4C. For determination of nonspecific binding, this incubation medium was supplemented with a 100-fold excess of unlabeled sAPPrec. After washing in ligand binding buffer, the cells were lysed in 0.5 M perchloric acid. Aliquots were used to determine radioactivity (Gamma 5.500; Beckmann Instruments, Munich, Germany) and amount of total DNA (Burton 1956). Whereas the saturation of binding sites is the main result obtained from the radiometric analysis of sAPPrec binding, the analysis of the specificity of binding requires confirmation by additional techniques such as surface plasmon resonance spectroscopy.
Surface Plasmon Resonance Spectroscopy
Protein-protein and cell-protein interactions were analyzed by surface plasmon resonance (SPR) spectroscopy as described (Malmqvist 1993). The Biacore biosensor (Uppsala, Sweden) uses SPR to probe the refractive index changes in a flow cell caused by the binding of molecules or cells to an immobilized ligand (Johnsson et al. 1991; Quinn et al. 1997). The Sensorchip CM5 and amine coupling reagents were provided by the manufacturers. Carboxymethylated dextran was activated by injection of 35 μl 1:1 mixture of N-hydroxysuccinimide (100 mM) and N-ethyl- N'-(dimethylaminopropyl)-carbodi-imide (4 M) at a flow rate of 5 μl/min. sAPPrec in sodium acetic buffer (10 mM) at a pH of 4.5 was immobilized, giving approximately 8000 response units. Excess reactive groups were saturated by the use of 35 μl ethanolaminehydrochloride (1 M, pH 8.5). B104 cells were preincubated for 15 min with different concentrations of sAPPrec in PBS. The flow rate was adjusted to 5 μl/min.
Statistical Analysis of Quantitative Data
Fluorescence intensity assays were performed by measuring a minimum of 40 cells per concentration. Fluorescence intensity assays were repeated three times in each experiment. The radiometric determinations were derived from two separate experiments and each point of concentration was completed in triplicate, i.e., data from three different wells were collected and analyzed statistically. Results were expressed as the mean ± SD. Data were analyzed for statistical differences by Student's unpaired t-test. p < 0.05 was considered statistically significant.
Results
Detection of Exogenously Added sAPPrec
Previous attempts to immunocytochemically visualize cell surface-bound sAPP failed because of the simultaneous recognition of the N-terminal region of APP by any of the available antibodies. Therefore, we developed antibodies against sAPPrec that were highly specific because no interference with endogenous APP in cell lysates or with secreted sAPP in culture media was observed (Figure 1). For immunocytochemical detection of endogenous APP, formaldehyde-fixed FRTL-5, B104, or PC 12 cells were permeabilized and incubated with an antibody against the C-terminal portion of APP. As shown in Figure 2 (insets in Figures 2A, 2C and 2E), the cells exhibited a strong perinuclear crescent-shaped staining pattern that has been shown to co-localize with staining of mannosidase II and therefore to be detectable mainly in the Golgi complex (Graebert et al. 1995a). The staining of the Golgi complex with the rabbit antiserum 2189 recognizing the C-terminus was the same compared to experiments in which the antibody 22C11 directed against the N-terminus was used (not shown), indicating that the C- and N-terminal regions are co-localized and suggesting that APP in the Golgi complex, at least in part, still carried the N-terminal portion of sAPP.
Selective detection of sAPPrec by antibody 3329 as shown by immunoblotting (Lane 5). Note that neither cellular APP in cell lysates (Lane 2) nor sAPP in the culture medium (Lane 4) was recognized by antibody 3329. For comparison, the detection of endogenous APP in cell lysates (Lane 1) or of sAPP released into the medium (Lane 3) is shown by the use of antibody 22C11.
Immunofluorescence microscopic visualization of exogenously added sAPPrec and of endogenous APP. Cell surface labeling with sAPPrec is shown in FRTL-5 (A), B104 (C), and PC 12 (E) cells using antibody 3329. (Insets) Localization of endogenous APP (antibody 22C11) is mainly observed in the Golgi complex. The co-localization of endogenous APP (red, antibody 2189) and of cell surface-bound sAPPrec (green, chicken antibody B1) is shown in FRTL-5 (B), B104 (D), and PC 12 (F) cells using confocal laser scanning microscopy. Note that exogenously added sAPPrec is assembled on the cell surfaces in the form of microdomains and that cell surface binding of sAPPrec can be clearly distinguished from the predominant Golgi staining for endogenous APP. Bars: A,C,E, insets = 10 μm; B,D,F = 5 μm.
For surface labeling and subsequent visualization, cells were incubated with sAPPrec at concentrations of 0.06-100 nM. Application of the anti-sAPPrec antibody (antiserum 3329) resulted in the clear visualization of cell surface-bound sAPPrec by both fluorescence and transmission electron microscopy (Figures 2 and 3). Labeling was observed on all free accessible cell surface domains, except for the domains adhering to the culture dish or the coverslips. The same result was obtained with the chicken anti sAPPrec antibody B1. Highest fluorescence intensity was observed at concentrations of 10 and 100 nM sAPPrec. Weakest but still detectable immunofluorescence was seen at 1 nM. It has been shown that the concentration of human sAPP in the circulation amounts to about 60 pM sAPP (Van Nostrand et al. 1991), which is below cytochemical detectability. We would like to point out, however, that this concentration is significant only for cells directly exposed to the bloodstream and may have no direct relevance for the concentration of sAPP in the intercellular spaces after its autocrine release.
Distribution of cell surface-bound sAPPrec shown by the use of antibody 3329 followed by immunogold labeling and silver enhancement. Note the labeling in the form of patches (brackets) 0.5-1.0 μm in diameter. Electron micrograph. Bar = 1 μm.
In cells incubated at 4C with sAPPrec followed by detection with anti-sAPPrec antibodies, sAPPrec was found to be clustered (Fig. 2) and to form patches of 0.5-1.0 μm in diameter on the cell surfaces (Fig. 3). These patches were evenly distributed over the cell surfaces of FRTL-5, B 104, and PC12 cells, excluding the membrane areas attached to the culture dish. This characteristic cell surface labeling pattern differed significantly from the labeling pattern of endogenous APP, as shown by double immunofluorescence labeling experiments (Figures 2B, 2D, and 2F). In these double labeling experiments, the simultaneous visualization of cellular APP and of exogenously added sAPPrec was made possible by the use of rabbit antiserum 2189 directed against the APP C-terminus and the chicken antibody B1 against sAPPrec. The use of the chicken antiserum B1 alone was previously tested and was observed to provide the same immunocytochemical staining as the rabbit anti-sAPPrec antibody 3329. The localization of cell surface-bound sAPPrec, as analyzed by confocal laser scanning microscopy, was compared with the results obtained with an antibody against His tag or with streptavidin-Cy3. The results indicated that the antibodies against sAPPrec provided the most sensitive detection technique as compared to the antibody against His tag or the vizualization of biotinylated sAPPrec with streptavidin-Cy3 (results not shown).
The labeling density of cell surface-bound sAPPrec differed markedly with the state of differentiation of the cells. Increased formation of neurites and of cell-cell interactions were taken as signs of cell differentiation. As shown in Figure 4 for B 104 cells, undifferentiated cells bound considerably more sAPPrec than fully mature ones.
Differentiation-dependent binding of sAPPrec to B104 cells as determined densitometrically. The binding of sAPPrec to differentiated cells is reduced (∗, p < 0.05). Bars represent SD.
Immunocytochemical Control Experiments (Not Shown)
Immunocytochemical staining was totally absent when the rabbit or chicken preimmune serum from nonimmunized animals was used or after omission of the primary antibody. This shows that the immunostaining was due to the presence of the corresponding antibodies and that nonspecific staining by the secondary fluorescently labeled antibody was negligible. Immunostaining was almost completely blocked when chicken or rabbit antiserum was preadsorbed with purified sAPPrec, thus confirming the specificity of the immunostaining with chicken and rat antisera.
Characteristics of sAPPrec Binding
Cells were incubated at 4C with [125 I]-sAPPrec at concentrations ranging from 0.1 nM to 100 nM. The results showed that saturation of binding appeared to be reached at 10 nM sAPPrec. A 100-fold excess of unlabeled sAPPrec showed almost complete inhibition of binding, thus indicating that sAPPrec binding was specific (Figure 5A).
(A) Saturation of sAPPrec binding to FRTL-5 cells. Maximal binding of [125 I]-sAPPrec was reached at concentrations of 10 nM (▪, total binding). Binding was strongly abolished by unlabeled sAPPrec used at 100-fold higher concentrations above iodinated sAPPrec. Δ, nonspecific binding; ∘, specific binding. (B) Competitive inhibition of sAPPrec binding to B104 cells as determined by surface plasmon resonance spectroscopy. Note that free sAPPrec at 4 μM almost completely inhibits binding of cells to immobilized sAPPrec. The binding intensity is expressed in response units (RU).
To ensure the specificity of sAPPrec binding to cell surfaces, B 104 cells were preincubated in suspension with different concentrations of sAPPrec at 4C. After this preincubation, their ability to bind to immobilized sAPPrec was monitored by surface plasmon resonance spectroscopy. Preincubation with 1 μM sAPPrec resulted in a decrease of binding by 60%. Binding was almost completely inhibited at 4 μM (Figure 5B) thus confirming that the sAPPrec binding was specific.
Influence of Heparinase I on sAPPrec Binding
Several growth factors such as FGF are known to bind to heparin (Schlessinger et al. 1995). We investigated the influence of cell surface associated heparin on the binding of sAPPrec to FRTL-5 cells, because sAPP also carries heparin-binding sites. As shown in Figure 6, the removal of cell surface-associated heparin by treatment of FRTL-5 cells with heparinase I did not affect the binding of sAPPrec.
Effect of heparinase I treatment on sAPPrec binding as quantitated by densitometric determination. sAPPrec binding was not affected by pretreatment of cells with heparinase I, suggesting that cell surface-associated heparin is not essential for sAPPrec binding to cells. Bars represent SD.
Sensitivity of sAPPrec Binding Towards Aldehyde Fixation
To further investigate the binding characteristics of sAPPrec to cell surfaces, FRTL-5 or PC 12 cells were fixed for 30 min with 4% paraformaldehyde before the incubation with sAPPrec was performed. Quantitation of bound sAPPrec was performed by densitometry as decribed in Materials and Methods. After fixation the binding of sAPPrec to cell surfaces was significantly reduced by 60-75% as shown in Figure 7.
Sensitivity of sAPPrec binding to FRTL-5 or PC 12 cells to pretreatment of cells with 4% formaldehyde (∗, p<0.05). Bars represent SD.
Discussion
The binding of sAPP to cell surfaces is of particular importance because it is the first step in the cascade of signal transduction events that lead to epithelial proliferation or to distinct physiological responses in other cell types. However, the visualization of exogenously added sAPP and its binding on cells is greatly hampered due to the ubiquitous expression of APP. The indirect localization of binding sites reported here makes use of recombinant sAPP (sAPPrec) and of the following features of sAPPrec: It elicits signal transduction and proliferation in thyrocytes (Pietrzik et al. 1998) comparable to endogenous sAPP (Popp et al. 1996), thereby suggesting that sAPPrec might bind to the same receptor. This is supported by recent observations showing that sAPPrec binding to cell surfaces is strongly inhibited by endogenous, epithelial cell derived sAPP from conditioned media (unpublished results). Despite this functional homology of both proteins, the structural differences of sAPPrec to endogenous sAPP are sufficient to allow the preparation of specific antibodies against sAPPrec. The ability of the antibodies to distinguish between sAPPrec and cellular APP or sAPP is presumably based on the lack of glycosylation which for sAPP is known to involve the formation of N- as well as O-bound carbohydrate side chains (Schubert et al. 1989; Pahlsson et al. 1992; Graebert et al. 1995b).
The anti-sAPPrec antibodies allowed selective localization of cell surface-bound sAPPrec in a variety of cell types in which highest fluorescence intensity appeared to be reached at 10 nM sAPPrec. Radiometric observations with iodinated sAPPrec showed that the binding of sAPPrec to its putative receptor is saturable reaching saturation at 10 nM sAPPrec. The specificity of binding was also shown by these radiometric determinations and confirmed by surface plasmon resonance spectroscopy in that binding of cells to immobilized sAPPrec was almost completely inhibited at 4 μM of free sAPPrec.
sAPP occurs in the circulation at concentrations of about 60 pM. However, platelets which are the main source of circulating sAPP may contribute up to 30 nM sAPP (Van Nostrand et al. 1991). This is well within the range of concentrations at which saturation of cell surface binding sites with sAPPrec occurs and at which cell surface binding can be detected using the technique described here. Reliable detection of sAPP binding was not observed below 1 nM sAPPrec. We would like to point out, however, that the reported normal concentration in the circulation of 60 pM does not necessarily correspond to the concentrations in the intercellular spaces into which sAPP is released in an autocrine fashion. This concentration is mainly regulated by the secretory release of sAPP which in thyrocytes has been shown to be stimulated by TSH (Graebert et al. 1995a). It may also be influenced by a number of other factors such as binding of sAPP to cell surfaces, endocytic removal of sAPP, the action of proteases on the cell surface (Brix et al. 1996), and sAPP adsorption to extracellular matrix constituents (Beyreuther et al. 1996). The actual concentration of free sAPP in the intercellular spaces is as yet unknown but may be higher than in the circulation.
The efficiency of sAPPrec specific antibodies for immunocytochemical detection of sAPPrec was compared to a number of other procedures which involved biotinylated sAPPrec or sAPPrec carrying a His tag on its C-terminal portion. Both sAPPrec derivatives, which are also able to bind to cell surfaces and to induce physiological responses such as cell proliferation, can be recognized by a specific antibody against the His tag or by fluorescently labeled streptavidin, thereby offering additional possibilities for the selective visualization of sAPPrec. Our observations show that the antibodies against sAPPrec provide the most sensitive detection device as compared to both other techniques.
As also shown in this study, the putative receptor for sAPPrec is highly sensitive to aldehyde fixation because the ability of cells to bind sAPPrec is markedly reduced after this treatment. Because of the preferential reaction of aldehyde groups with lysine residues (Hopwood 1972), this inhibitory effect on sAPPrec binding points to the proteinaceous nature of sAPPrec binding sites. Apparently, the number of binding sites for sAPPrec is not constant because it appears to be dependent, e.g., in the neuronal cell line B 104, on the state of differentiation. We take this observation as further support for the existence of a cell membrane receptor for sAPPrec. It has been suggested, however, that sAPP might bind to cell surface associated heparin and that this binding assists the receptor to come into contact with sAPP (Kounnas et al. 1995). Nevertheless, binding remains unaffected by treatment of cells with heparinase, suggesting that cell surface associated heparin is not the prime factor for sAPPrec binding to cell surfaces.
sAPPrec binding occurs in the form of patches evenly distributed on the free accessible cell surfaces. These patches are preformed, i.e., they are present at 4C and not induced by ligand binding. They appear to correspond to membrane rafts that have been implicated in participation in cellular processes such as sorting in polarized cells (Lisanti et al. 1989; Brown and Rose 1992; Kurzchalia et al. 1992) or signal transduction and cell adhesion (Harder et al. 1998; for review see Simons and Ikonen 1997; Hooper 1998) and which appear to be involved in the processing of APP in hippocampal neurons (Simons et al. 1998). Glycosylphosphatidyl-inositol (GPI)-anchored proteins have been reported to be a general feature of membrane proteins assembled in cell surface microdomains (Friedrichson and Kurzchalia 1998; Varma and Mayor 1998). It will therefore be an interesting aspect of future experiments to see whether sAPPrec binding sites are complexed to GPI-anchored proteins or whether they themselves are GPI-anchored proteins. It will be equally interesting to dissect the role of the sAPPrec binding microdomains and to determine their structural constituents, i.e., their protein and lipid composition.
In conclusion, we present a method for selective visualization of cell surface-bound sAPPrec without interference by cellular APP or sAPP. The method provides the experimental basis to indirectly locate cell surface binding sites that appear to be assembled in cell surface microdomains.
Footnotes
Acknowledgements
Supported by Deutsche Forschungsgemeinschaft (SFB 284), by a Nordrhein Westfalen research grant, and by Fonds der Chemischen Industrie.
We thank Drs K. Beyreuther and G. Multhaup for the monoclonal antibody 22C11 and W. Neumüller for discussions and critical reading of the manuscript. We are grateful to Ms Babette Baumann and Ms. Beate Wolf for technical assistance, to Ms Andrea Roth for typing the manuscript, and to Ms Ann Icking for assistance throughout the binding experiments.