Chondroitin sulfate proteoglycan form of the Alzheimer's beta-amyloid ...

Communication

THEJOURNAL OF BIOLOGICAL CHEMISTRY Vol. 267, No. 20, Issue of July 15, pp. 13819-13822.1992 0 1992 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

Chondroitin Sulfate Proteoglycan Form of the Alzheimer’s0Amyloid Precursor* (Received for publication, April 9, 1992)

Junichi Shioi, John P. Anderson, James A. Ripellino, and Nikolaos K. RobakisS From the Department of Psychiatry andFishberg Research Center for Neurobiology, Mount SinaiSchool of Medicine, New York, New York 10029

The Alzheimer’s amyloid B protein is derived from a family of membrane glycoproteins termed amyloid precursor proteins (APP). Here we show that APP exists as the core protein of a chondroitin sulfate (CS) proteoglycan, ranging in apparent molecular size from 140 to 250 kDa, secreted by glial cell line C6. After partial purification on ion-exchange and gel chromatography, the secreted APP proteoglycan was recognized on Western blots by several antibodies specific to different regions of APP. Chondroitinase AC or ABC treatment of our samples completely eliminated the high molecular weight proteoglycan with a concomitant increase in the APP protein. This digestedproduct reacted with an anti-stub antibody which recognizes 4-sulfated disaccharide. Sequencing of the N terminus of the core protein of this CS proteoglycan yielded 18 residues identical to the N terminus sequence of the mature APP. Quantitative analysis showed that, in this cell line, about 90%of the secreted nexin I1 form of APP occurs in the proteoglycan form, suggesting that the CS chains have a role in the biological function of this protein. The close proximity of two consensus CS attachment sites to both the N terminus of the amyloid ,8 protein and the secretase cleavage site, suggests that the CS chains may affect the proteolysis of APP and production of the amyloid /3 protein.

extracellular portions of two of the APP isoforms contain a 56-amino acid region with highsequence homology to the Kunitz-type serine protease inhibitors (KPI) (7-9). Following glycosylation, APPs are cleaved by APPsecretase at the extracellular region, close to the transmembrane sequence, and secreted (10,ll). Nexin 11, the secreted APP form derived from the KPI containing precursors (12, 13), migrates on SDS-PAGE asa 120-kDa protein detectedby antisera specific to the KPI insert(14). Chondroitin sulfateproteoglycans (CSPG) consistof a core protein towhich one or more chondroitin sulfate (CS)glycosaminoglycan (GAG) chains arecovalently attached. Secreted, membrane-bound, and intracellulargranule CSPG have been detected. They are involved in a variety of cellular functions including cell adhesion and migration, cell-cell communication, and modulation of growth factor activities (for review see Refs. 15 and 16). MATERIALS ANDMETHODS

Cell Cultures-Confluent cultures of rat neuroglial C6 cells (17) were split 1:20 and grown in Dulbecco’s modified Eagle’s medium (GIBCO) supplemented with 10% fetal bovine serum (JRH Biosciences). For preparation of conditionedmedium for proteoglycan purification, C6 cells were washed with Dulbecco’s modified Eagle’s medium 2 days after passage, and then the culture was continued in serum-free N2medium (18). Serum-freeconditioned medium was collected up to 14 days after passage. Purification of CSPG Form of APP-C6 cell-conditioned medium was applied to a dextran sulfate-Sepharose column(19). The column was washedwithbufferA (20 mM Tris, 1.0 mM EDTA, pH 8.2) containing 0.15 M NaCl, and the bound material was eluted with buffer A containing 1.0 M NaC1. The fraction eluted from the dextran sulfate column was made 4 M in guanidinium isothiocyanate, 10 mM dithiothreitol (DTT), 170 mM P-mercaptoethanol (&ME), and 0.1 mM phenylmethylsulfonyl fluoride. The samplewas applied toa column of Sephacryl S-400 HR equilibrated in 5.0 M guanidinium isothiocyanate, 50 mM Tris, 0.1 mM EDTA, 5 mM DTT, pH 8.0, and eluted with the same buffer. Fractions were analyzed by immunoblotting. Substitution of guanidine hydrochloride for guanidine isothiocyanate gave similar results. Samples from the dextran sulfate columnwere dialyzed against 20 mM Tris-buffered saline and 6 M urea, pH 8.2. After dialysis, the samples were made 10 mM in DTT and 1%in p-ME, applied to a Amyloid fibrils in Alzheimer’s disease accumulate in the DEAE-Sepharosecolumn,and washedextensivelywith thesame neuritic plaque cores and cerebral blood vessels. The major buffer. The bound materialswere then elutedwith a gradient of 0.15component of the amyloid depositions is the 4-kDa amyloid 1.0 M NaCl in the samebuffer. Chondroitinase Digestion-Twenty-microliter aliquots of the conprotein (APP,’ A4 peptide (1, 2)) which is part of at least three larger membrane glycoproteins termed amyloid precur- ditioned medium were incubated with chondroitinaseABC (0.1 unit/ sor proteins (APPs) (3-9). The three APP proteins contain ml) from Proteusvulgaris (Sigma) orwith chondroitinase AC (0.0013 (Sigma) asdescribed (20) in the presence of protease inhibbetween 695 and 770 residues and consist of a large extracel- unit/ml) itors (2 mM phenylmethylsulfonylfluoride, 10 pg/ml pepstatin, 10 lular region, a transmembrane domain of about 25 residues, pg/ml leupeptin, 20 pg/ml aprotinin, and 1 mM 1,lO-phenanthroline) and a small cytoplasmic portion of 47 amino acids. The in a final volume of 30 pl for 4 h at 37 “C. S D S Electrophoresis and Zmmunoblots-Samples were made 3% in * This work was supported by National Institute on Aging Grants SDS and 5% in @-ME,placed in boiling water for 5 min, and then AGO8200 and AG05138. The costs of publication of this article were electrophoresed on an 8% polyacrylamide minigel (Bio-Rad) in the defrayed in part by the payment of page charges. This article must presence of 0.1% SDS (21). Proteins were electroblotted to polyvitherefore be hereby marked “advertisement” in accordance with 18 nylidene difluoride membrane (Immobilon, Millipore) and then imU.S.C. Section 1734 solely to indicate this fact. munodetected with anti-APPantiseraas described (14, 22). For $ T o whom correspondence shouldbe addressed: MountSinai detection of core protein by anti-stub antibodies, thefollowing monoSchool of Medicine, Dept. of PsychiatryandFishbergResearch clonal antibodies (Seikagaku Co. Rockville, MD) were employed at Center for Neurobiology, One Gustave Levy Place, Box 1229, New 2000 times dilution:Adi-0-sulfate,Adi-4-sulfate, and Adi-6-sulfate York, NY 10029. Tel.: 212-241-9380; Fax: 212-831-1947. directed against non-sulfated,4-sulfated, and 6-sulfated disaccharide ’ The abbreviations used are: ApP, amyloid p protein; APP, amy- “stub”units, respectively (23).Forquantification,densitometric loid precursor protein; KPI, Kunitz-type serine protease inhibitor; measurements were performed with a computerized Magiscan JoyceSDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electropho- Loebl image analysis system. resis; CS, chondroitin sulfate; CSPG, chondroitin sulfate proteoglyFingerprints of Enzymatic and Chemical Digests-C6 nexin I1 or can; GAG, glycosaminoglycan; DTT, dithiothreitol; p-ME, 0-mercap- proteoglycan,purified by chromatographyondextransulfateand toethanol. DEAE as described above, or chondroitinase-digested proteoglycan

13819

Proteoglycan Form of APP

13820

was denatured by 0.1% SDS and 5 min of boiling, made 0.6% in noctyl glucoside, and then incubatedwith 0.05 pg of endoprotease Lysc (sequence grade, Sigma) in 30 pl of 0.1 M ammonium bicarbonate, pH 7.7, a t 37 “C for 4 h (24). For formic acid treatment, partiallypurified C6 nexin I1 and nexin 11-free proteoglycan were incubated with 70% formic acid a t room temperature for 24 h (25). For cyanogen bromide treatment, partially purified C6 nexin I1 (3.7 pg) or nexin 11-free proteoglycan (2.5 pg of total protein) was suspended in200 pl of 70% formic acidcontaining 2 pg of tryptophan. Then 20 pg of cyanogen bromide (Sigma) in 20 p1 of 70% formic acid was added to each sample, and the tubes were sealed with parafilm after argonflushing(26). Tubes were kept in the dark at room temperature for 24 h. Samples were then dilutedwith 800 pl of water and lyophilized in a Speed-Vacsystem (Savant).Dried materials were washed twice with 1 ml of water by centrifuge lyophilization. Dried materials were solubilized in sample buffer containing SDS and pM E (21). All digested samples were analyzed by immunoblotting. Amino Acid Sequencing-APP proteoglycan was partially purified and completely separated from nexin I1 by chromatography on dextran sulfate and DEAE-Sepharose and, after dialysis in 0.1 M Tris maleate, 0.06 M sodium acetate, pH 6.0, digested with chondroitinase ABC as described above. The digest was separated by SDS-PAGE and transferred topolyvinylidene difluoride (Immobilon PSEQ, Millipore)(24).Thebandcontainingthe 120-kDacore protein was excised and sequenced in a Problott cartridge (Applied Biosystems). Sequencing was preformed by theProteinSequencingCenter at SUNY Health Science Center at Brooklyn. Edman degradation and phenylthiohydantoin-derivativeanalysis were carried out on a model 470 gas-phase protein sequenator, model 990A controller data processor, connected on-line to a model 120A microbore high pressure liquid chromatography phenylthiohydantoin-aminoacid analyzer, with the results quantitated using a model 900 data processing module (all of the above equipment was fromApplied Biosystems). The initial yield was 20.5 pmol, and theaverage repetitive yield was 92.4%. Only one predominant aminoacid was obtained in eachsequencing cycle.

RESULTS

A

a

b

c

d

a

b

c Dd ae b

c

I t

2 0 5 - n

205116-

I

coo(.

d

116- . I 97-:

”

0

66-

9 7-

a

66-

20511697-

b

c

d

e

f

g

h

b4b414’

i

-

- 205 -116

-

97 66

45

j

’

m

66-

FIG. 1. Chondroitinase sensitivity and purification of the APP proteoglycan from C6 cell-conditioned medium. A , 2O-pl aliquots of the C6 cell-conditioned medium were incubated without (lane a ) or with (lane b ) 0.003 units of chondroitinase ABC in a final volume of 30 pl for 4 h a t 37 “C. Samples were analyzed as described under “Materials and Methods” using 8% polyacrylamide gel and anti-R7 antiserum. Lanes c and d , a lighter exposure of lanes Q and b, respectively, in a preflashed film. Densitometric measurements were performed on these lanes. B, APP, partially purified from C6 cell culture medium on a dextran sulfatecolumn ( l o n e a ) was fractionated by gel exclusion chromatography in 5.0 M guanidinium isothiocyanate, as described under “Materials and Methods.” Lanes b-e show serial 1.5-ml fractions, spanning 16.0 ml ( b ) through 22.0 ml ( e ) of eluent. Other fractions, before or after, showed no anti-APP reactivity. Void volume was 13 ml, and total volume was 28 ml. Samples were electrophoresed on an 8% polyacrylamide gel and probed with R7 antiserum. C, samples from the dextran sulfate column (Fig. lB, lane a ) were fractionated by DEAE chromatography, as described under “Materials andMethods.” Samples from the collected fractions were electrophoresed on a 10% polyacrylamide gel and probed with R7 antiserum. Salt concentration increases from fraction Q to j . D, partially purified proteoglycan from C6 cell culture medium (lanes Q and b) or pure nexin I1 isolated from PC12 cell culture medium (22) (lanes c and d ) were incubated without (lanes Q and c) or with (lanes b and d ) chondroitinase ABC. Samples were electrophoresed on an 8% polyacrylamide gel and probed with R7 antiserum. Numbers are molecular masses of standard proteins, inkDa.

Conditioned media from a rat C6 glioma cell line (17) were analyzed on Western blots stained with the R7 antiserum, specific for the KPI insert of the APP. This antiserum has been used extensively forthe characterizationof nexin I1 (14, 22). Fig. LA (lune a) illustratesthatinSDS-PAGE gels, besides the 120-kDa nexin 11, additional diffuse staining, similar to the staining pattern associated with proteoglycans still reactive toward R7 antiserum (Fig. lC, lunes e-j). Treat(27, 28), was detected by the R7 antiserum in the region of ment of a proteoglycan-enriched fraction with chondroitinase 140-250 kDa. Incubation of the culture medium with chon- ABC resulted in thedegradation of the R7 antiserum-reactive droitinase ABC, an enzyme that degrades both CS and der- proteoglycan and a largeincreasein the nexin I1 signal, matan sulfateGAG chains (20), resulted in the disappearancesuggesting that nexin I1 was the core protein of this proteoof the diffuse reactivity and a large increase in the levels of glycan (Fig. 1D). nexin I1 (Fig. L4, lanes b and d). These data suggested that Although it has been reported that nexin I1 binds to GAG chondroitinase ABC removed the GAG portionfromthe chains like heparin, this non-covalent association is rather proteoglycan and generated a core protein which on SDS- weak since nexin I1 is eluted from a heparin column in the PAGE displayed the same mobility and immunoreactivity as presence of 0.3-0.6 M NaCl (30). The failure of strong denanexin 11. Densitometric measurement of the signals before turing agents such as guanidinium, urea, or boiling in the and after chondroitinase digestion (Fig. LA, lunes c and d ) presence of 3.0% SDS and5% P-ME, performed before samshowed that free nexin I1 represented about 10% while the ples were loaded on the Laemmli gel, to dissociate the R7 proteoglycan nexin I1 was about 90% of the total secreted antiserum-reactive antigen from the high molecular weight nexin 11. proteoglycan provided strong evidence for a covalent bond T o further elucidate the nature of this diffuse reactivity, between nexin I1 and theGAG chains. the R7 antiserum-reactive proteoglycan waspartially purified Since it is still possible that thecore protein of this proteofrom culturemedium of C6 cellsby binding to dextran sulfate-glycan is distinct from nexin I1 although it shares the same Sepharose followed by elution with 1.0 M NaCl and subse- mobility and R7 antiserum immunoreactivity with the latter, quent gel filtration in the presence of 4 M guanidinium iso- nexin 11-free proteoglycan was digested with chondroitinase thiocyanate or guanidinium hydrochloride. Under these con- ABC and thenprobed either with the GID antiserum directed ditions nexinI1 was retained longer in thecolumn relative to against aminoacids 175-186 of the APP(30) or with the R47 t h e R7 antiserum-reactive proteoglycan (Fig. 1B). The latter antiserum directed againstamino acids 1-17 of the APP displayed the size heterogeneitywhich is typical of many sequence. Fig. 2 shows that both antiserarecognized the core proteoglycans and is attributed to variations in the number protein produced afterchondroitinase digestion, while no of GAG chains attached to the core protein, the length of nexin I1 was detected before digestion. Furthermore, the moeach GAG chain, or both(29). DEAE chromatography in the bility of the core prqtein was identical to themobility of nexin presence of 6 M urea resulted in the separationof free nexin 11. Similarresults were obtainedwith R5 antiserum (14) I1 (Fig. lC, lunes a-d) from the proteoglycan. The latter was directed against APP amino acids 476-491 (numbering ac-

13821


nase digestion, however, the level of nexin I1 increased, and the resultant core protein also reacted with the anti-chon205droitin 4-sulfate antibody (Fig. 3B, lane f), clearly indicating that theproteoglycan-derived nexin I1 contained disaccharide 11697GAG units attached to it. No reaction was observed either with the anti-chondroitin 6-sulfate antibody or the antibody 66directed against the non-sulfated “stub” sugar. These results 45indicated that the chondroitin GAG units proximal to the FIG. 2. Reactivity of the chondroitinase ABC digests of the nexin I1 core protein contained 4-sulfated residues and proNexin-free vided strong evidence of the proteoglycan nature of nexin 11. proteoglycan with GID and R47 anti-APP antisera. C6 proteoglycan was incubatedwithout (lanes a and c) or with (lanes Structural evidence for the identity of the core protein of b and d ) chondroitinase ABC. Lane e, pure PC12 nexin I1 (see Fig. 1D).Samples were electrophoresedon an 8%polyacrylamide gel and this proteoglycan was obtained after treatment of both puriprobed with GID antiserum (lanes a and b), R47 antiserum (lanes c fied nexin I1 and proteoglycan with endoprotease lysine C (24), formic acid, which cleaves peptides at Asp-Pro bonds and d ) , or R7 antiserum (lane e ) . (25), or cyanogen bromide (26). Fig. 4 shows that treatment of nexin 11-free proteoglycan with each of these agents proA a ‘b a cb d e t g h duced fragments displaying an identical mobility and immuFnoreactivity with the fragments produced from proteoglycan! a 205 free nexin 11. To confirm the identity of the core protein, 205nexin 11-free proteoglycan was digested with chondroitinase 116 ABC, and the resulting 120-kDa core protein was processed and sequenced as described under “Materials and Methods.” 11666 The first18cycles yielded the sequence LEVPTDGNAGLLA97EPQIA, which corresponds to the N-terminal sequence of FIG. 3. A, nexin 11-free proteoglycan was incubated without (lane APP after removal of the leader peptide (33). The structural a) or with (lane b ) chondroitinase AC (0.0013 unit/ml) (Sigma). data combined with the immunological data clearly demonSamples were electrophoresedon a 6% polyacrylamide gel and then strate that the core protein of the detected proteoglycan is probed with R7 antiserum. B, a proteoglycan/nexin I1 mixture was nexin 11. incubated without (lanes u, c, e, and g) or with (lanes b, d, f, and h) a

b

c

d

e

..

chondroitinase ABC, electrophoresed on an 8%gel, and then probed either with R7 antisera (lanes a and b ) or with the following monoclonal antibodies: Adi-0-sulfate (lanes c and d ) , Adi-4-sulfate (lanes e and f), and Adi-6-sulfate (lanes g and h). Numbers represent molecular mass markers in kDa.

DISCUSSION

The demonstration that nexin I1 can occur as a CSPG suggests new directions for the efforts to elucidate the biological function of the APP proteins. Additional results in our laboratory indicate that a fraction of the cell surface fullcording to Ref. 8). It should be noted that theR47 antiserum length APP also contains GAG chains? Cell-associated or failed to recognize the intactproteoglycan (Fig. 2).Decreased secreted proteoglycans are involved in important biological immunoreactivity of intact proteoglycans relative to theircore functions including cell adhesion, migration, and growth conproteins is often observed and may be due to the masking of trol (15, 16). They have been shown to play a critical role in epitopes by the GAG chains (27, 28). Since it is very unlikely several basement membrane-related diseases including aththat two unrelated proteins display identical mobilities and erosclerosis (34) and metastasis (35). Cell surface APP has immunoreactivities for four antisera, each directed against a been reported to promote neuronal cell adhesion (36). This distinct epitope, these data provided further support that effect may be mediated by the CSPG form of APP, which has nexin I1 is the core protein of this proteoglycan. In agreement also been found in the brain? Dysfunction of this form of with this conclusion, R1 antiserum, directed against the cyAPP may affect neuronal adhesion, thus contributing to the toplasmic portion of APP (14), failed to react with the chonneuronal degeneration observed in Alzheimer’s disease. droitinase digest (data notshown). Secreted small size proteoglycans inhibit cell adhesion by Since the enzyme chondroitinase ABC will digest both CS interfering with the binding of cell membrane-associated proand dermatan sulfate chains, nexin 11-free proteoglycan was teoglycans to theextracellular matrix (15,16).In vitro, CSPGs treated with chondroitinase AC, which is specific for CS GAG inhibit axon initiation and elongation, thus modulating neural chains (20). Fig. 3A shows that thisenzyme digested the GAG patterning (37, 38). This raises the possibility that the seand yielded a core protein with a mobility identical to thatof creted APP proteoglycan regulates cell adhesion and/or nexin 11, demonstrating that CS GAG chains are attached to neural patterning by a similar mechanism. nexin 11. Furthermore, treatment of the proteoglycan with The proteoglycan nature of the amyloid precursor may have heparinase (EC 4.2.2.7) or heparitinase I (EC 4.2.2.8), which important implications for the production of the ADP, which degrades heparin or heparin and heparan sulfate GAGS, reseems to be the result of aberrant APP processing (39). The spectively (31,32), failed to produce a shift in mobility of R7 GAG chains of CSPGs are attached to the Ser residue of a antiserum immunoreactivity of the proteoglycan (data not Ser-Gly dipeptide motif (15,16,40). APPhas only three such shown). sequences, at positions 57,637, and 660. The last two consenCore proteins of CSPGs are specifically detected by monosus GAG attachment sites closely flank Asp-653, the N terclonal antibodies directed against the highly immunogenic minus of the ADP peptide; indeed, ser-660 is part of the ADP unsaturated disaccharide “stubs” that remain attached to the sequence, corresponding to ADP amino acid 8, and is also core protein after digestion of the GAG chains with chonlocated only 9 residues upstream of the APP secretase cleavdroitinase ABC. Antibodies specific for unsulfated, 4-sulfated, age site (22). Inhibition of cleavage at this site has been or 6-sulfated “stubs” arealso used to determine the natureof implicated in the production of the ADP (39). Given the large the sulfate residues of the GAG disaccharide proximal to the core protein (23). Fig. 3B ( l a n e a) illustrates that before * J. Shioi, L. M. Refolo, S. Efthimiopoulos, and N. K. Robakis, chondroitinase digestion, a proteoglycan fraction containing manuscript in preparation. free nexin I1 reacted only with R7 antisera. After chondroitiJ. A. Ripellino and N. K. Robakis, unpublished observations.


13822 a

A 205-

b

c

.

d

e

f

B

L.

b

c

d

c

a

., b

t

L .

1169766- , 45 -

a

11697-

-116 97

-

n @ - 6 6

29-

664529-

? -

-

45 29

FIG. 4. Fingerprints of enzymatic andchemical digests of proteoglycan from C6 cell culture medium. A, partially purified C6 nexin I1 (0.6 pg of total protein; lanes a and b), proteoglycan (0.6 pg; lanes d and f), or chondroitinase-digested proteoglycan (0.6 pg; lanes c and e) was incubated without (lanes a, e, and f ) or with (lanes b, c, and d) 0.05 pg of endoprotease Lys-C (sequence grade, Sigma) as described under “Materials and Methods.” After digestion, samples were analyzed by SDS-PAGE on a 12% polyacrylamide gel and then probed with R7 antiserum. The proteoglycan failed to penetrate this gel due to the high percent of acrylamide (lane f). B, partially purified C6 nexin I1 (1.5 and 3.7 pg; lanes a and b, respectively) and nexin 11-free proteoglycan (2.5 and 1.0 pg; lanes c and d, respectively) were incubated without ( a and d ) or with (lanes b and c) 70% formic acid at room temperature for 24 h. Samples were electrophoresed on a 9% polyacrylamide gel and then probed with GID antiserum. Note that only a fraction of nexin I1 or proteoglycan was cleaved by formic acid. C,partially purified C6 nexin I1 (3.7 pg; lane a) or nexin 11-free proteoglycan (2.5 pg; lane b) was processed for cyanogen bromide digestion as described under “Materials and Methods.” Samples were electrophoresed on a 9% polyacrylamide gel and probed with GID antiserum. Nostrand, W. E., Wagner, S. L., Suzuki, M., Choi, B. H., Farrow, J. charge and size of the GAG chains (15, 16), attachment of 13. Van S., Geddes, J. W., Cotman, C. W., and Cunningham,D. D. (1989)Nature GAG on serine 637 or 660 could affect the proteolytic cleav341,546-548 L. M., Salton, S. R. J., Anderson, J. P., Mehta, P., and Robakis, N. age(s) thatproduce the APP by, for example, hindering access 14. Refolo, K. (1989)Biochem. Biophys. Res. Commun. 164,664-670 of a clipsin-like protease (41)to Asp-653. 15. Ruoslahti, E. (1988)Annu. Reo. Cell Biol. 4, 229-255 R.L., Busch, S. J., and Cardin, A. D. (1991)Physiol. Reo. 71, Proteoglycans are degraded in the lysosomes to progres- 16. Jackson, 481-539 sively smaller fragments by proteolysis of the core protein 17. Benda, P., Lightbody, L., Sato, G., Levine, L., and Sweet,W. (1968)Science 161,370-371 and glycolytic removal of the GAG chains (42). It is expected 18. Bottenstein, J. E., and Sato, G. H. (1979)Proc. Natl. Acad. Sci. U. S. A. that APPproteoglycan is degraded through the samemecha76,514-517 Van Nostrand, W. E., and Cunningham, D. E. (1987)J. Biol. Chem. 262, 19. nism and may be the precursor of the potentially amyloido8508-8514 genic fragments detected in lysosomes (43). APP mutations 20. Yamagata, Y., Saito, H., Habuchi, O., and Suzuki, S. (1968)J. Biol. Chem. 243,1523-1535 (44,45) ordysfunction of a glycolytic enzyme could interfere 21. Laemmli, U. K. (1970)Nature 227,680-685 with the normal removal of the GAG residues and alter the 22. Anderson, J. P., Esch, F. S., Keim, P. S., Sambamurti, K., Lieberburg, I., and Robakis, N. K. (1991)Neurosci. Lett. 128,126-128 proteolytic processing of APP. Thus, the demonstration that 23. Caterson, B., Calabro, T., and Hampton, A. (1987)in Biology of ProteoglyAPP molecules contain CS chains suggests new potential cam (Wight, T. N., and Mecham, R. P., eds) pp. 1-16,Academic Press, New York lesion sites that could lead to aberrant APP processing and 24. Matsudaira, P. (1990)Methods Enzyrnol. 182,602-613 production of APP. 25. Landon, M. (1977)Methods Enzymol. 47, 145-149

Acknowledgments-Protein sequencing was performedbyDr. J. Rushbrook and L. Siconolfi-Baez of the Protein Sequencing Center Department of Biochemistry, SUNY Health Science, Brooklyn, NY. We thank Y. Chen for excellent technical assistance, Drs. T. Saitoh and P. Mehta for GID and R47 antisera, respectively, and Dr. L. Refolo for helpful discussions. REFERENCES 1. Glenner, G. G., and Wong, C. W. (1984)Biochem. Biophys. Res. Commun.

120,885-890 2. Masters, C. L., Simms, G., Weinman, N. A,, Multhaup, G., McDonald, B. L., and Beyreuther, K. (1985)Proc. Natl. Acad. Sci. U.S. A. 82, 4245A944

3. Kt&,“J., Lemaire, H.-G., Unterbeck, A., Salbaum, J. M., Masters, C. L., Grzeschik. K.-H.. Multhauo. G.. Bevreuther. K.. andMuller-Hill. B. (1987)N&re 325; 733-738’ ’ 4. Robakis, N. K., Ramakrishna, N., Wolfe, G., and Wisniewski,H. M. (1987) Proc. Natl. Acad. Sci. U. S. A. 84.4190-4194 5. Goldgaber, D., Lerman, M. I., McBhde, W. O., Saffiotti, U., and Gajdusek, D. C. (1987)Science 235,877-880 6. Tanzl, R. E., Gusella, J. F., Watkins, P. C., Bruns, G. A. P., St. GeorgeHyslop, P., Van Keuven, M. L., Patterson, D., Pagan, S., Kurnit, D. M., and Neve, R. L. (1987)Science 235,880-884 7. Kitaguchi, N., Takahashi, Y., Tokushima, Y., Shiojiri, S., and Ito,H. (1988) Nature 331,530-532 8. Ponte, P., Gonzalez-DeWhitt, P., Schilling, J., Miller, J., Hsu, D., Greenberg, B., Davis, K., Wallace, W., Lieberburg, I., Fuller, F., and Cordell, B. (1988)Nature 33,525-527 9. Tanzi, R. E., McClatchey, A. I., Lamperti, E. D., Villa-Komaroff, L., Gusella, J. F., and Neve, R. L. (1988)Nature 331,528-530 10. Sisodia, S. S., Koo, E. H., Beyreuther, K., Unterbeck, A. J., and Price, D. L. (1990)Science 248,492-495 11. Esch, F. S., Keim, P. S., Beattie, E. C., Blacher, R. W., Culwell, A. R., Oltersdorf, T., McClure, D., and Ward, P. J. (1990)Science 248, 1122,

~

.”.

I

~

1l 9 A

12. Oltersdorf, T., Fritz, L. C., Schenk, D. B., Lieberburg, I., Johnson-Wood, K. L., Beattie, E. C., Ward, P. J., Blacher, R. W., Dovey, H. F., and Sinha, S. (1989)Nature 341, 144-147

26. Godinot, C., Colorio, S., Cretin, F., Incaurgarat, B., Deleage, G., and Roux, R. (1989)Biochimie 71,917-929 21. Gowda, D. C., Hogue-Angeletti, R., Margolis, R. K., and Margolis, R. U. (1990)Arch. Riochem. Bzophys. 281,219-224 28. Rauch, U., Gao, P., Janetzko, A., Flaccus, A,, Hilgenberg, L., Tekotte, H., Margolis, R. K., and Margolis, R. U. (1991)J. Biol. Chem. 266, 1478514801 29. Kjellen, L., and Lindahl, C. (1991)Annu. Rev. Biochem. 60,443-475 30. Schubert, D., LaCorbiere, M., Saitoh, T., and Cole, G. (1989)Proc. Natl. Acad. Sci. U. S. A. 86,2066-2069 31. Linker, A,, and Hovingh, P. (1972)Methods Enzymoj. 28,902-911 32. Silva, M. E., and Dletrlch, C. P. (1974)Biochem. Btophys. Res. Commun. 56,965-972 33. Palmert, M. R., Podlisny, M. B., Witker, D. S., Oltersdorf, T.,Younkin, L. H., Selkoe, D. J., and Younkin,S. G. (1989)Proc. Natl.Acad. Sei. U. S. A. 86,6338-6342 34. Wagner, W. D., Salisbury, B. G. J., and Rowe, H. A. (1986)Arteriosclerosis 6,407-417 35. Newman, S. A., Frenz, D.A., Hasegawa, E., and Akiyama, S. K. (1987) Proc. Natl. Acad. Sci. U. S. A. 84.4791-4795 36. Breen, K. C., Bruce, M., and Anderton, B. H. (1991)J. Neurosci. Res. 28, 90-100 37. Snow, D. M., Lemmon, V., Carrino, D. A,, Caplan, A.I., and Silver, J. (1990)Exp. Neurol. 109,111-130 38. Brittis, P. A., Canning, D. R., and Silver, J. (1992)Science 255,733-736 39. Robakis, N. K., Sambamurti, K., Anderson, J. P.. Mehv, P., and Refolo, L.M. (1991) in Growth Factors and Alzhermer’s DLsease (Heftl, F., Brachet, P., Will, B., and Christen,Y., eds) pp.208-215,Sprmger-Verlag, Berlin 40. Antonsson, P., Heinegard, D., and Oldberg, A. (1989)J. Biol. Chem. 264, 16170-16173 41. Nelson, R. B., and Siman, R. (1990)J. Biol. Chem. 265,3836-3843 42. Iozzo, R. V. (1987)J. Riol. Chem. 262,1888-1900 43. Golde, T. E., Estus, S., Younkin, L. H., Selkoe, D. J., and Younkin, S, G. (1992)Science 255,728-730 44. Goate, A., Chartier-Harlin, M. C., Mullan? M., Brown, J., Crawford, F., Fidani, L., Giuffra, L., Haynes, A,, Irvlnt N., James, L., Mant, R., Newton, P., Rooke, K., Roques, P., Talbot, ., Perxak-Vance, M., Roses, A., Williamson, R., Rossor, M., Owen, M., and Hardy, J. (1991)Nature 349,704-706 45. Murrell, J., Farlow,M., Ghetti, B., and Benson, M. D. (1991)Science 254, 97-99