The Transmembrane Aspartates in Presenilin 1 and 2 Are Obligatory for γ-Secretase Activity and Amyloid β-Protein Generation
2000; Elsevier BV; Volume: 275; Issue: 5 Linguagem: Inglês
10.1074/jbc.275.5.3173
ISSN1083-351X
AutoresW. Taylor Kimberly, Weiming Xia, Talat Rahmati, Michael S. Wolfe, Dennis J. Selkoe,
Tópico(s)Trace Elements in Health
ResumoThe discovery that a deficiency of presenilin 1 (PS1) decreases the production of amyloid β-protein (Aβ) identified the presenilins as important mediators of the γ-secretase cleavage of β-amyloid precursor protein (APP). Recently, we found that two conserved transmembrane (TM) aspartates in PS1 are critical for Aβ production, providing evidence that PS1 either functions as a required diaspartyl cofactor for γ-secretase or is itself γ-secretase. Presenilin 2 (PS2) shares substantial sequence and possibly functional homology with PS1. Here, we show that the two TM aspartates in PS2 are also critical for γ-secretase activity, providing further evidence that PS2 is functionally homologous to PS1. Cells stably co-expressing TM Asp → Ala mutations in both PS1 and PS2 show further accumulation of the APP-derived γ-secretase substrates, C83 and C99. The production of Aβ is reduced to undetectable levels in the conditioned media of these cells. Furthermore, endoproteolysis of the exogenous Asp mutant PS2 is absent, and endogenous PS1 C-terminal fragments are diminished to undetectable levels. Therefore, the co-expression of PS1 and PS2 TM Asp → Ala mutants suppresses the formation of any detectable PS1 or PS2 heterodimeric fragments and essentially abolishes the production of Aβ. These results explain the residual Aβ production seen in PS1-deficient cells and demonstrate the absolute requirement of functional presenilins for Aβ generation. We conclude that presenilins, and their TM aspartates in particular, are attractive targets for lowering Aβ therapeutically to prevent Alzheimer's disease. The discovery that a deficiency of presenilin 1 (PS1) decreases the production of amyloid β-protein (Aβ) identified the presenilins as important mediators of the γ-secretase cleavage of β-amyloid precursor protein (APP). Recently, we found that two conserved transmembrane (TM) aspartates in PS1 are critical for Aβ production, providing evidence that PS1 either functions as a required diaspartyl cofactor for γ-secretase or is itself γ-secretase. Presenilin 2 (PS2) shares substantial sequence and possibly functional homology with PS1. Here, we show that the two TM aspartates in PS2 are also critical for γ-secretase activity, providing further evidence that PS2 is functionally homologous to PS1. Cells stably co-expressing TM Asp → Ala mutations in both PS1 and PS2 show further accumulation of the APP-derived γ-secretase substrates, C83 and C99. The production of Aβ is reduced to undetectable levels in the conditioned media of these cells. Furthermore, endoproteolysis of the exogenous Asp mutant PS2 is absent, and endogenous PS1 C-terminal fragments are diminished to undetectable levels. Therefore, the co-expression of PS1 and PS2 TM Asp → Ala mutants suppresses the formation of any detectable PS1 or PS2 heterodimeric fragments and essentially abolishes the production of Aβ. These results explain the residual Aβ production seen in PS1-deficient cells and demonstrate the absolute requirement of functional presenilins for Aβ generation. We conclude that presenilins, and their TM aspartates in particular, are attractive targets for lowering Aβ therapeutically to prevent Alzheimer's disease. presenilin amyloid β-protein Chinese hamster ovary β-amyloid precursor protein N-terminal fragment C-terminal fragment transmembrane enzyme-linked immunosorbent assay wild type polyacrylamide gel electrophoresis immunoprecipitation N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine Mutations in presenilin 1 (PS1)1 and presenilin 2 (PS2) are responsible for the most severe and common form of dominant, familial Alzheimer's disease (1.Selkoe D.J. Trends Cell Biol. 1998; 8: 447-453Abstract Full Text Full Text PDF PubMed Scopus (808) Google Scholar). These missense mutations have been shown to subtly alter the proteolytic processing of the β-amyloid precursor protein (APP). APP is cleaved by α- or β-secretase, and the membrane-retained 83- and 99-residue C-terminal fragments produced by these cleavages (C83 and C99, respectively) serve as the substrates for a second cleavage by a protease designated γ-secretase. The products of the C99 cleavage include the 40- and 42-residue amyloid β-peptides (Aβ), which accumulate to high abundance in myriad amyloid plaques in the brain. The familial Alzheimer's disease mutations in the presenilins appear to selectively increase the production and deposition of Aβ42 in cell culture (2.Scheuner D. Eckman C. Jensen M. Song X. Citron M. Suzuki N. Bird T.D. Hardy J. Hutton M. Kukull W. Larson E. Levy-Lahad E. Viitanen M. Peskind E. Poorkaj P. Schellenberg G. Tanzi R. Wasco W. Lannfelt L. Selkoe D. Younkin S. Nat. Med. 1996; 2: 864-870Crossref PubMed Scopus (2282) Google Scholar, 3.Xia W. Zhang J. Kholodenko D. Citron M. Podlisny M.B. Teplow D.B. Haass C. Seubert P. Koo E.H. Selkoe D.J. J. Biol. Chem. 1997; 272: 7977-7982Abstract Full Text Full Text PDF PubMed Scopus (272) Google Scholar, 4.Tomita T. Maruyama K. Saido T.C. Kume H. Shinozaki K. Tokuhiro S. Capell A. Walter J. Grunberg J. Haass C. Iwatsubo T. Obata K. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 2025-2030Crossref PubMed Scopus (355) Google Scholar), in the brains of transgenic mice (5.Duff K. Eckman C. Zehr C., Yu, X. Prada C.-M. Perez-Tur J. Hutton M. Buee L. Harigaya Y. Yager D. Morgan D. Gordon M.N. Holcomb L. Refolo L. Zenk B. Hardy J. Younkin S. Nature. 1996; 383: 710-713Crossref PubMed Scopus (1325) Google Scholar, 6.Borchelt D.R. Ratovitski T. van Lare J. Lee M.K. Gonzales V. Jenkins N.A. Copeland N.G. Price D.L. Sisodia S.S. Neuron. 1997; 19: 939-945Abstract Full Text Full Text PDF PubMed Scopus (901) Google Scholar, 7.Citron M. Westaway D. Xia W. Carlson G. Diehl T. Levesque G. Johnson-Wood K. Lee M. Seubert P. Davis A. Kholodenka D. Motter R. Sherrington R. Perry B. Yao H. Strome R. Lieberburg I. Rommens J. Kim S. Schenk D. Fraser P. St. George-Hyslop P. Selkoe D.J. Nat. Med. 1997; 3: 67-72Crossref PubMed Scopus (1169) Google Scholar), and in the brains of familial Alzheimer's disease patients (8.Lemere C.A. Lopera F. Kosik K.S. Lendon C.L. Ossa J. Saido T.C. Yamaguchi H. Ruiz A. Martinez A. Madrigal L. Hincapie L. Arango L.J.C. Anthony D.C. Koo E.H. Goate A.M. Selkoe D.J. Arango V.J.C. Nat. Med. 1996; 2: 1146-1148Crossref PubMed Scopus (436) Google Scholar). Furthermore, in vitroevidence demonstrates that Aβ42 is more likely than Aβ40 to aggregate and polymerize into fibrils (3.Xia W. Zhang J. Kholodenko D. Citron M. Podlisny M.B. Teplow D.B. Haass C. Seubert P. Koo E.H. Selkoe D.J. J. Biol. Chem. 1997; 272: 7977-7982Abstract Full Text Full Text PDF PubMed Scopus (272) Google Scholar, 9.Jarrett J.T. Berger E.P. Lansbury Jr., P.T. Biochemistry. 1993; 32: 4693-4697Crossref PubMed Scopus (1768) Google Scholar). These Aβ42 fibrils appear to act as a seed for subsequent plaque formation (1.Selkoe D.J. Trends Cell Biol. 1998; 8: 447-453Abstract Full Text Full Text PDF PubMed Scopus (808) Google Scholar). The mechanism by which the presenilins cause this selective increase in Aβ42 has remained a mystery until recently. The discovery that a deficiency in PS1 inhibits the normal processing of APP and reduces both Aβ40 and Aβ42 production suggested that PS1 facilitates γ-secretase activity (10.De Strooper B. Saftig P. Craessaerts K. Vanderstichele H. Gundula G. Annaert W. Von Figura K. Van Leuven F. Nature. 1998; 391: 387-390Crossref PubMed Scopus (1560) Google Scholar). The region of function of PS1 in this regard has been further localized to two aspartates within the putative 6th and 7th transmembrane (TM) domains, which are required for Aβ production (11.Wolfe M.S. Xia W. Ostaszewski B.L. Diehl T.S. Kimberly W.T. Selkoe D.J. Nature. 1999; 398: 513-517Crossref PubMed Scopus (1699) Google Scholar). The mutation of either aspartate 257 (TM6) or aspartate 385 (TM7) to alanine interferes with both the endoproteolysis of PS1 and the generation of Aβ. This occurs by a dominant negative mechanism that replaces endogenous PS1 N- and C-terminal fragments (NTFs and CTFs) with the nonfunctional mutant holoprotein. However, in neurons that lack PS1, or in Chinese hamster ovary (CHO) cells that overexpress PS1 Asp → Ala mutant protein, residual Aβ production is still present (10.De Strooper B. Saftig P. Craessaerts K. Vanderstichele H. Gundula G. Annaert W. Von Figura K. Van Leuven F. Nature. 1998; 391: 387-390Crossref PubMed Scopus (1560) Google Scholar, 11.Wolfe M.S. Xia W. Ostaszewski B.L. Diehl T.S. Kimberly W.T. Selkoe D.J. Nature. 1999; 398: 513-517Crossref PubMed Scopus (1699) Google Scholar). This remaining peptide, which is approximately 20–40% of the normal Aβ level, is postulated to arise from the residual PS2 activity in these cells. To test this hypothesis, we generated Asp → Ala mutations at the equivalent residues in PS2, aspartates 263 and 366. We first showed that these PS2 aspartyl mutants, like their PS1 aspartyl mutant counterparts, have a similar effect on APP processing, providing further evidence that PS2 is functionally homologous to PS1. We then stably introduced the PS2 D366A construct into CHO cells overexpressing human PS1 D257A and wild-type human APP. We found that the presence of both PS1 and PS2 aspartyl mutants completely abolishes γ-secretase processing of APP. We observed additional accumulation of the C83 and C99 APP CTFs and a suppression of total Aβ production to levels that were undetectable by a sensitive sandwich ELISA. This absence of Aβ corresponded with an absence of detectable PS1 and PS2 endoproteolytic fragments. We conclude that functional PS1 or PS2 heterodimers are absolutely required to generate Aβ and that the TM aspartates in particular are critical in both proteins. Full-length wild-type (wt) PS2 cDNA was used as the template to introduce D263A or D366A by site-directed mutagenesis, and the genes were inserted into the vector pcDNA3.1(+) containing either the zeocin or hygromycin resistance genes (Invitrogen). The identities of the mutant genes were confirmed by DNA sequencing. For transfection, plasmids (10 μg) were introduced into various CHO cell lines in 10-cm dishes using LipofectAMINE (Life Technologies, Inc.). Plasmids encoding wt PS1, wt PS2, and PS1 D385A have been characterized (11.Wolfe M.S. Xia W. Ostaszewski B.L. Diehl T.S. Kimberly W.T. Selkoe D.J. Nature. 1999; 398: 513-517Crossref PubMed Scopus (1699) Google Scholar, 12.Kim T.-W. Pettingell W.H. Hallmark O.G. Moir R.D. Wasco W. Tanzi R.E. J. Biol. Chem. 1997; 272: 11006-11010Abstract Full Text Full Text PDF PubMed Scopus (202) Google Scholar). Chinese hamster ovary (CHO) cells stably overexpressing the 751-residue human wt APP751 (7W cells (13.Koo E.H. Squazzo S. J. Biol. Chem. 1994; 269: 17386-17389Abstract Full Text PDF PubMed Google Scholar); designated APPwt here) were maintained in 200 μg/ml G418 (Life Technologies, Inc.). Cells stably co-expressing human wt APP together with either wt PS2 (designated PS2wt (3.Xia W. Zhang J. Kholodenko D. Citron M. Podlisny M.B. Teplow D.B. Haass C. Seubert P. Koo E.H. Selkoe D.J. J. Biol. Chem. 1997; 272: 7977-7982Abstract Full Text Full Text PDF PubMed Scopus (272) Google Scholar)) or D257A PS1 (designated 2-1, (11.Wolfe M.S. Xia W. Ostaszewski B.L. Diehl T.S. Kimberly W.T. Selkoe D.J. Nature. 1999; 398: 513-517Crossref PubMed Scopus (1699) Google Scholar)) were maintained in 200 μg/ml G418 plus 250 μg/ml Zeocin (Invitrogen). Stable CHO cells co-expressing wt APP, wt PS1, and wt PS2 (designated PS1+2wt) were maintained in 200 μg/ml G418, 250 μg/ml Zeocin, and 2.5 μg/ml puromycin (Sigma). Stable CHO cells co-expressing wt APP, D257A PS1, and D366A PS2 (designated 2A-2, 2A-19, and 2A-21) were maintained in 200 μg/ml G418, 250 μg/ml Zeocin, and 250 μg/ml hygromycin B (Life Technologies, Inc.). Polyclonal antibody C7 to APP732–751 (APP751 numbering) was described previously (11.Wolfe M.S. Xia W. Ostaszewski B.L. Diehl T.S. Kimberly W.T. Selkoe D.J. Nature. 1999; 398: 513-517Crossref PubMed Scopus (1699) Google Scholar, 14.Podlisny M.B. Tolan D. Selkoe D.J. Am. J. Pathol. 1991; 138: 1423-1435PubMed Google Scholar). Monoclonal antibodies 13G8 and 8E5 were gifts of P. Seubert and D. Schenk and are directed against APP732–751 (13G8) or APP500–648 (8E5). Polyclonal antibody 4627 was raised against PS1457–467 and monoclonal antibody 13A11 recognizes PS1294–309 (15.Podlisny M.B. Citron M. Amarante P. Sherrington R. Xia W. Zhang J. Diehl T. Levesque G. Fraser P. Haass C. Koo E.H.M. Seubert P. St. George-Hyslop P. Teplow D.B. Selkoe D.J. Neurobiol. Dis. 1997; 3: 325-337Crossref PubMed Scopus (274) Google Scholar). Antibody PS2L was a gift of T. Iwatsubo (4.Tomita T. Maruyama K. Saido T.C. Kume H. Shinozaki K. Tokuhiro S. Capell A. Walter J. Grunberg J. Haass C. Iwatsubo T. Obata K. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 2025-2030Crossref PubMed Scopus (355) Google Scholar) and recognizes PS2316–339. Polyclonal antibody 2C20 recognizes the last 20 residues of the PS2 C terminus. 2972, which recognizes the PS2 N terminus (residues 2–84), was a gift of C. Haass (16.Walter J. Grunberg J. Schindzielorz A. Haass C. Biochemistry. 1998; 37: 5961-5967Crossref PubMed Scopus (57) Google Scholar). For experiments in which [35S]methionine labeling was performed, cells were starved for 30 min in media lacking methionine and fetal bovine serum, followed by a 1-h label in [35S]methionine-containing media. In all experiments, cells were solubilized in a lysis buffer containing 50 mm Tris, pH 7.6, 150 mm NaCl, 2 mm EDTA, 1% Nonidet P-40 and a protease inhibitor mixture (5 μg/ml leupeptin, 5 μg/ml aprotinin, 2 μg/ml pepstatin A, and 0.25 mm phenylmethylsulfonyl fluoride (Sigma)). Lysates were adjusted to equalize total protein concentration using the BCA protein assay (Pierce). Samples were precleared with protein A-Sepharose at 4 °C for 30 min and then precipitated with primary antibody and protein A-Sepharose for 2 h. Immunoprecipitates were washed for 20 min at 4 °C in the above lysis buffer containing 500 mm NaCl and then in lysis buffer with 0.1% SDS. This was followed by a final wash in lysis buffer. Samples were eluted in 2× sample buffer (20% glycerol, 4% SDS, 10% β-mercaptoethanol) at 65 °C for 5 min, resolved by SDS-PAGE on 10, 14, or 4–20% Tris glycine or Tris-Tricine gels (Novex), and transferred to polyvinylidene difluoride (Millipore). Western blotting using ECL Plus detection was performed according to the manufacturer's instructions (Amersham Pharmacia Biotech). For experiments in which quantitative analysis was performed, blots were directly scanned using a Storm 840 scanner (Molecular Dynamics) with a chemiluminescence detection setting. Individual bands were quantitated using ImageQuant software (Molecular Dynamics). Statistical significance was determined using the Student's t test (two-tailed). Aβ sandwich ELISAs were performed as described (17.Johnson-Wood K. Lee M. Motter R. Hu K. Gordon G. Barbour R. Khan K. Gordon M. Tan H. Games D. Lieberburg I. Schenk D. Seubert P. MConlogue L. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 1550-1555Crossref PubMed Scopus (585) Google Scholar). The capture antibody was 266 (to Aβ residues 13–28). The reporter antibody was biotinylated 3D6 (to Aβ residues 1–5). Values were expressed as a fraction of Aβ produced by the wt cell lines (APPwt, PS2wt, and PS1+2wt). The absence of PS1 in neuronal cells decreases the γ-secretase cleavage of APP (10.De Strooper B. Saftig P. Craessaerts K. Vanderstichele H. Gundula G. Annaert W. Von Figura K. Van Leuven F. Nature. 1998; 391: 387-390Crossref PubMed Scopus (1560) Google Scholar). We recently reported that mutation of either of the two transmembrane (TM) aspartates of PS1 induces a similar suppression of Aβ generation in cell culture (11.Wolfe M.S. Xia W. Ostaszewski B.L. Diehl T.S. Kimberly W.T. Selkoe D.J. Nature. 1999; 398: 513-517Crossref PubMed Scopus (1699) Google Scholar). Based on this evidence, we hypothesized that PS1 is either γ-secretase itself or a cofactor for γ-secretase. PS1 and PS2 share significant homology (18.Levy-Lahad E. Wasco W. Poorkaj P. Romano D.M. Oshima J. Pettingell H., Yu, C. Jondro P.D. Schmidt S.D. Wang K. Crowley A.C. Fu Y.-H. Guenette S.Y. Galas D. Nemens E. Wijsman E.M. Bird T.D. Schellenberg G.D. Tanzi R.E. Science. 1995; 269: 973-977Crossref PubMed Scopus (2241) Google Scholar, 19.Rogaev E.I. Sherrington R. Rogaeva E.A. Levesque G. Ikeda M. Liang Y. Chi H. Lin C. Holamn K. Tsuda T. Mar L. Sorbi S. Nacmias B. Piacentini S. Amaducci L. Chumakov I. Cohen D. Lannfelt L. Fraser P.E. Rommens J.M. St. George-Hyslop P.H. Nature. 1995; 376: 775-778Crossref PubMed Scopus (1801) Google Scholar), and it has been suggested that they may possess redundant roles with respect to APP processing. We therefore mutated the equivalent TM aspartate residues in PS2 to alanines in order to ascertain the effect on the accumulation of C83 and C99, the APP C-terminal substrates for γ-secretase. Mammalian expression vectors containing the cDNAs for wild-type (wt) human PS1, wt human PS2, D385A PS1, D263A PS2, or D366A PS2 were transiently transfected into Chinese hamster ovary (CHO) cells stably expressing the 751-residue wt human APP (APPwt cell line). Cell lysates were collected and normalized for total protein content, followed by immunoprecipitation with a C-terminal specific polyclonal APP antibody (C7) and Western blotting with a C-terminal specific monoclonal antibody (13G8). The expression of the PS1 aspartyl mutant or either of the PS2 aspartyl mutants resulted in an increase of C83 and C99 when compared with cells expressing wt PS1 or PS2 (Fig.1). Because mutation of the transmembrane aspartates in both PS1 and PS2 had similar effects on APP catabolism, we investigated the effects of expressing both mutant proteins together. We transfected a PS2 D366A cDNA into CHO cells that overexpress wt APP and PS1 D257A (cell line 2-1) to generate clones that stably express all three human proteins. We screened for expression of the mutant PS2 holoprotein by Western blotting using antibody 2972, which recognizes human but not hamster PS2 with high affinity (Fig. 2 a). Clone 2A-2 demonstrated a high expression level of mutant PS2 holoprotein as indicated by a pair of bands at ∼50–55 kDa (Fig. 2 a, lane 4) that co-migrated with the PS2 bands in the previously characterized wt PS2-expressing CHO cell line (Fig. 2 a, lane 2 and Ref. 3.Xia W. Zhang J. Kholodenko D. Citron M. Podlisny M.B. Teplow D.B. Haass C. Seubert P. Koo E.H. Selkoe D.J. J. Biol. Chem. 1997; 272: 7977-7982Abstract Full Text Full Text PDF PubMed Scopus (272) Google Scholar). These bands were not present in CHO cells lacking human PS2 (Fig. 2 a, lanes 1 and 3). To confirm the expression levels of APP, we performed specific IP/Western analysis on several wt APP-overexpressing cell lines. Using antibody C7 to immunoprecipitate APP and antibody 8E5 to Western blot, the expression levels of full-length human APP were demonstrated to be identical among the cell lines (Fig. 2 b, upper). Importantly, APP expression in 2A-2 is closely similar to its parental cell line, 2-1 (Fig. 2 b, lanes 4 and 5). To confirm that the cells had similar rates of APP production, we [35S]methionine-labeled cells for 1 h and then immunoprecipitated full-length APP with antibody C7. Fig. 2 b(lower) demonstrates equivalent levels of immatureN-glycosylated APP (N-APP) and matureN-,O-glycosylated APP (N+O-APP) in all cell lines. We also verified the expression status of PS1 holoprotein. Fig.2 c demonstrates equivalent levels of expression of human PS1 holoprotein among a wt PS1-expressing cell line (PS1wt) and cell lines 2-1 and 2A-2 (lanes 2, 4, and 5). This expression is compared with APPwt and PS2wt, which contain only endogenous PS1 holoprotein (lanes 1 and 3). Therefore, 2A-2 has levels of human wt APP and PS1 D257A comparable to its parental cell line 2-1 and also expresses high levels of mutant PS2 holoprotein. To assess the combined effects of the aspartyl mutants on APP catabolism, we performed IP/Western analysis of the C83 and C99 APP fragments. To compare appropriately the levels of these fragments, we also detected the full-length APP from the same samples by blotting the top half of the filter with 8E5. To control for the effects of expressing multiple genes in a given cell, we compared the C83 and C99 levels among three wt-containing lines: wt APP (APPwt), wt APP, and wt PS2 (PS2wt), and wt APP, wt PS1, and wt PS2 (PS1+2wt). The single, double, or triple stable expression of these cDNAs in CHO cells resulted in approximately similar levels of C83 and C99 (Fig. 3 a, lanes 1–3). Therefore, the stable transfection of three genes (and the effect of the selection drugs required to maintain them) did not have any effect on APP proteolytic events that generate these fragments. Additionally, we studied the maturation of APP using fine pulse-chase time points to determine if there were a more subtle but nonspecific perturbation in protein metabolism. Following a 5-min pulse and analyzing full-length APP every 5–15 min thereafter (to 60 min), we found no difference in the rate of N-,O-glycosylation of APP when it was expressed at endogenous levels versus in double transfectants versus in triple transfectants. 2W. Xia, W. T. Kimberly, and D. J. Selko, unpublished observations. Therefore, the overexpression of three genes does not appear to have any effect on APP processing. Likewise, there was no alteration of the SDS-PAGE profile of total [35S]methionine-labeled proteins among the three kinds of transfectants (not shown). In cells that stably express the aspartate mutant PS proteins, however, we observed a markedly increased accumulation of C83 and C99. 2-1 cells, which express high levels of PS1 D257A, exhibit such an increase (Fig. 3 a, lane 4), as reported (11.Wolfe M.S. Xia W. Ostaszewski B.L. Diehl T.S. Kimberly W.T. Selkoe D.J. Nature. 1999; 398: 513-517Crossref PubMed Scopus (1699) Google Scholar). 2A-2 has further elevated levels (Fig. 3 a, lane 5), a result that is very clearly demonstrated when the cell lysates are diluted 1:100 and then subjected to IP (Fig. 3 a, lanes 9 and 10). At this dilution, only the predominant C83 species is visible in the aspartate mutant cell lines. To determine whether these mutations were interfering with α- or β-secretase processing of APP, we looked at the secretion of APPs and found that it was not inhibited in the aspartate mutant cell lines, consistent with our previous report (Ref.11.Wolfe M.S. Xia W. Ostaszewski B.L. Diehl T.S. Kimberly W.T. Selkoe D.J. Nature. 1999; 398: 513-517Crossref PubMed Scopus (1699) Google Scholar; and data not shown). Thus, the combined expression of PS1 D257A and PS2 D366A induces further accumulation of the C83 and C99 fragments, as compared with PS1 D257A alone. We also investigated the effects of two additional clones that co-express PS1 and PS2 aspartate mutations, 2A-19 and 2A-21. We used phosphorimaging analysis to determine quantitatively the level of the C83 fragments in these lines as well as in 2-1, 3-1, and 2A-2. We established a ratio of C83 to N-APP within each cell line and compared that ratio to the value obtained for 2-1. The three PS1 and PS2 aspartate mutant clones had mean relative ratios between 1.58 and 1.83, and their combined relative ratio was significantly greater than 2-1 (1.67 ± 0.08, p < 0.001, TableI). To validate this quantitative approach, we also measured the C83 to N-APP ratio in the previously characterized cell line 3-1 (11.Wolfe M.S. Xia W. Ostaszewski B.L. Diehl T.S. Kimberly W.T. Selkoe D.J. Nature. 1999; 398: 513-517Crossref PubMed Scopus (1699) Google Scholar). This cell line inhibits Aβ secretion to a similar extent as 2-1. As expected, there was no significant difference between 3-1 and 2-1 in relative C83 accumulation (0.99 ± 0.10, Table I), confirming the validity of this method.Table IIncreased accumulation of the C83 APP CTF in PS1 + PS2 aspartate mutant cell linesCell lineAsp → Ala mutant inFold increase in APP CTF C83 (mean ± S.E.)p valueaDifference from line 2-1 (Student's ttest).2-1PS11.003-1PS10.99 ± 0.102A-2PS1 + PS21.62 ± 0.23<0.052A-19PS1 + PS21.83 ± 0.532A-21PS1 + PS21.58 ± 0.26<0.05All 2A cell linesPS1 + PS21.67 ± 0.08<0.001The C83 to N-APP ratio was quantitatively determined within each cell line using ImageQuant software. The ratio for 2-1 was assigned a value of 1.00, and all other lines are expressed relative to that value. Cell line 3-1 (described previously (11.Wolfe M.S. Xia W. Ostaszewski B.L. Diehl T.S. Kimberly W.T. Selkoe D.J. Nature. 1999; 398: 513-517Crossref PubMed Scopus (1699) Google Scholar)) was included as a control that should not differ from 2-1 in relative C83 accumulation. The means represent three (2A-21) or four (3-1, 2A-2, 2A-19) independent determinations.a Difference from line 2-1 (Student's ttest). Open table in a new tab The C83 to N-APP ratio was quantitatively determined within each cell line using ImageQuant software. The ratio for 2-1 was assigned a value of 1.00, and all other lines are expressed relative to that value. Cell line 3-1 (described previously (11.Wolfe M.S. Xia W. Ostaszewski B.L. Diehl T.S. Kimberly W.T. Selkoe D.J. Nature. 1999; 398: 513-517Crossref PubMed Scopus (1699) Google Scholar)) was included as a control that should not differ from 2-1 in relative C83 accumulation. The means represent three (2A-21) or four (3-1, 2A-2, 2A-19) independent determinations. To investigate Aβ production, conditioned media were collected and analyzed by a sensitive and specific sandwich ELISA. Within each experiment, the values for the wt-expressing cell lines (APPwt, PS2wt, and PS1+2wt) were averaged and assigned a value of 1. The Aβ levels from all cell lines were then expressed as a fraction of that value. 2-1 showed a 64% reduction in Aβ relative to cells expressing wt presenilin (35.8 ± 5.4%, n = 6; Fig. 3 b), closely similar to our previously reported data on this line (11.Wolfe M.S. Xia W. Ostaszewski B.L. Diehl T.S. Kimberly W.T. Selkoe D.J. Nature. 1999; 398: 513-517Crossref PubMed Scopus (1699) Google Scholar). Strikingly, 2A-2 showed no detectable Aβ production ( 97% reduction (Fig. 3 b). Even when media were collected after long periods of conditioning (36 h), no Aβ was detected (data not shown). We did not measure Aβ42 levels since the total population of Aβ (which includes capture of the 42-amino acid peptide) was undetectable. Furthermore, the decrease in Aβ levels in conditioned media is not a result of impaired secretion since intracellular Aβ is undetectable in 2-1 cells, in contrast to wt-PS1-expressing cells (11.Wolfe M.S. Xia W. Ostaszewski B.L. Diehl T.S. Kimberly W.T. Selkoe D.J. Nature. 1999; 398: 513-517Crossref PubMed Scopus (1699) Google Scholar). Accordingly, microsomes isolated from cells expressing Asp mutant PS are unable to generate new Aβ, whereas wt PS-containing microsomes do so robustly (20.Xia, W., Wolfe, M., Ostaszewski, B. L., Rahmati, T., Zhang, J., and Selkoe, D. J. (1999) Soc. Neurosci. Abstr. 219.7Google Scholar). Therefore, the reduction in Aβ detected in the media of aspartate mutant PS-expressing cell lines reflects the specific inhibition of γ-secretase-mediated production. The stable expression of the PS1 aspartyl mutants inhibits the endoproteolytic cleavage of the exogenous protein and also suppresses the level of endogenous hamster PS1 fragments in a dose-dependent manner (11.Wolfe M.S. Xia W. Ostaszewski B.L. Diehl T.S. Kimberly W.T. Selkoe D.J. Nature. 1999; 398: 513-517Crossref PubMed Scopus (1699) Google Scholar). Thus, expression of the PS1 aspartyl mutants yields only full-length mutant holoprotein. We therefore investigated the accumulation of the PS2 NTF in cell line 2A-2 to determine if a similar phenomenon occurred. Cell lysates were resolved by SDS-PAGE, and PS2 NTFs were detected by Western blotting with the human-specific 2972 antibody. In the PS2wt and PS1+2wt cell lines, an ∼35-kDa NTF band was present (Fig.4 a, lanes 2 and 3). This band was not detected in either the wt APP or the 2-1 cell lines, which do not overexpress human PS2 (Fig. 4 a, lanes 1 and4). A human PS2 NTF was not present in the 2A-2 line, even though it expresses full-length protein at substantially higher levels than the other human PS2-expressing cells (Fig. 4 a, 5th lane). Even after long exposures, no band was detectable (data not shown). Importantly, co-expression of wt PS1 and wt PS2 does not abolish the formation of the PS2 NTF (Fig. 4 a, 3rd lane), even though PS2 holoprotein expression is not as high in the PS1+2wt cells as in 2A-2. This indicates that the lack of fragments in 2A-2 is due specifically to the mutation of the aspartate residues. We then attempted to detect human PS2 CTF with a different set of antibodies. By using polyclonal antibody PS2L for immunoprecipitation and polyclonal antibody 2C20 for Western blotting, we detected an ∼25-kDa band in PS2wt and PS1+2wt cells (Fig.4 b, lanes 2 and 3). Again, we failed to detect any PS2 CTFs in cell line 2A-2 (Fig. 4 b, lane 5). Therefore, the transmembrane aspartates of PS2 are critical for its endoproteolysis, consistent with the effect of the TM aspartates in PS1. The level of PS1 or PS2 fragments affects the relative heterodimer formation of the other (21.Thinakaran G. Harris C.L. Ratovitski T. Davenport F. Slunt H.H. Price D.L. Borchelt D.R. Sisodia S.S. J. Biol. Chem. 1997; 272: 28415-28422Abstract Full Text Full Text PDF PubMed Scopus (286) Google Scholar). Overexpression of PS1 can suppress the formation of endogenous fragments of PS2 and vice versa. Based on these findings, it was proposed that PS1 and PS2 compete for limiting cellular factors, which in turn allow endoproteolysis and stabilization of the fragments (21.Thinakaran G. Harris C.L. Ratovitski T. Dav
Referência(s)