inhibition of-amyloid formation identifies proteolytic precursors and

Neuron, Vol. 14, 651-659, March, 1995, Copyright © 1995 by Cell Press

Inhibition of -Amyloid Formation Identifies Proteolytic Precursors and SubceUular Site of Catabolism

Jeffrey Higaki, Diana Quon, Ziyang Zhong, and Barbara Cordell Scios Nova Incorporated 2450 Bayshore Parkway Mountain View, California 94043

Summary

Cerebral deposition of p-amyloid protein is a pathologi- cal feature central to Alzheimer's disease. Production of J]-amyloid by proteolytic processing of the p-amyloid precursor protein (pAPP) is a critical initial step in J~-amy- Ioidogenesis. We use an inhibitor of pAPP processing to block p-amyloid peptide formation. Application of the inhibitor to cultured cells results in an accumulation of proteolytic intermediates of pAPP, enabling a precursor-product relationship between pAPP carboxy-terminal fragments and ~-amyloid peptides to be demonstrated directly. In the presence of inhibitor, these amyloidogenic carboxy.-terminal fragments can be degraded to nonamyloidogenic products. The catabolism of pAPP carboxy-terminal intermediates and the formation of p-amyloid peptides are likely to in- volve an early endosomal compartment as the subcellular site of processing.

Introduction

Metabolism of 13-amyloid precursor protein (~APP) in AIz- heimer's disease and Down's syndrome, as well as some elderly normal individuals, leac:s to the production and accumulation of a 4 kDa proteolytic cleavage product, the 13-amyloid peptide. Two major pathways of ~APP processing have been identified that are implicated in the formation of ~-amyloid. One pathway involves proteolytic cleavages that result in the secretion of truncated I3APP and the concomitant formation of 8-12 kDa carboxy- terminal remnants (Caporaso e~ al., 1992; Golde et al., 1992; Estus et al., 1992; Haass et al., 1992a). The bulk of these carboxy-terminal secretion products are nonamy- Ioidogenic, since secretory cleavage most frequently occurs within the 13-amyloid domain situated in the carboxy- terminal region of the precursor protein (Esch et al., 1990). Only a small proportion of secretory cleavage products appears to be potentially amyloidogenic (Seubert et al., 1993). Evidence exists supporting both intracellular (Sam- bamurti et al., 1992; De Strooper et al., 1993; Kuentzel et al., 1993) and membrane surface sites (Sisodia, 1992; Haass et al., 1992a; Kuentzel et al., 1993) for the constitu- tive secretory cleavage process, and it appears that multiple locales of cleavage exist along the intracellular secretory pathway of ~APP.

A second processing pathway involving the endosomal/ lysosomal system has also been suggested as the site of formation for amyloidogenic 13APP carboxy-terminal fragment and J3-amyloid. This pathway was identified through

observations with cultured mammalian cells treated with inhibitors of lysosomal activity such as chloroquine, E64, or leupeptin. These compounds promote the accumulation of a number of ~APP carboxy-terminal derivatives, including larger potentially amyloidogenic fragments (Ca- poraso et al., 1992; Haass et al., 1992a; Golde et al., 1992; Hayashi et al., 1992; Busciglio et al., 1993; Siman et al., 1993). Further evidence for the participation of lysosomes in the generation of amyloidogenic processing products of J3APP comes from the presence of such fragments in crude lysosomal preparations from cultured cells (Haass et al., 1992a) and the association of active lysosomal enzymes with ~-amyloid deposits in Alzheimer's disease brain (Cataldo and Nixon, 1990).

Although lysosomal inhibitors have been shown to cause an accumulation of ~APP carboxy-terminal proteolytic cleavage intermediates, including those that could be processed further to yield J3-amyloid peptide, these inhibi- tore have little or no effect on the formation of the 4 kDa J3-amyloid peptide. Specifically, treatment of cell cultures with leupeptin or E64 does not block the formation of l~-amy- Ioid (Shoji et al., 1992; Haass et al., 1993; Busciglio et al., 1993). Short periods of treatment with chloroquine or NH4CI result in modest effects on ~-amyloid production (Busciglio et al., 1993); however, extended treatments are inhibitory (Shoji et al., 1992; Haass et al., 1993). These results indicate that the lysosome may not be participating directly in the processing of ~APP to ~-amyloid peptide. Additional pharmacological studies employing inhibitors of Golgi processing such as monensin and brefeldin A were found to block the formation of ~-amyloid, suggesting that the 4 kDa peptide is generated in the secretory pathway distal to the endopla:~mic reticulum and early Golgi (Busciglio et al., 1993; Haass et al., 1993). These data imply that there may be multiple subcellular sites of I3APP processing giving rise to a collection of amyloidogenic carboxy-terminal fragments, only a subset of which may be precursors of 13-amyloid peptide.

An inhibitor that acts directly on the processing of B-amy- Ioid would be useful in resolving which ~APP carboxy- terminal fragments are the authentic proteolytic precursors of ~-amyloid, as well as in defining the subcellular locale(s) involved in 13APP catabolism and ~-amyloidogenesis. Here we describe the application of an inhibitor capable of blocking 13-amyloid formation in vitro to address these issues.

Results

Inhibition of p-Amyloid Processing by the MDL 28170 Inhibitor Cultured mammalian cells have been reported to produce and secrete 13-amyloid peptides that include an intact 4 kDa ~-amyloid peptide identical to that purified from AIz- heimer's disease brain and an amino-terminally truncated 3 kDa peptide derived from the secretory cleavage of ~APP (Haass et al., 1992b; Shoji et al., 1992; Busciglio et al., 1993). Therefore, we utilized Chinese hamster ovary

Neuron 652

A kDa M 1 2 3 4 5 6 7 8

29

18

14

6

3

B kDa M 1 2 3 4 5 6 7 8

18

14

C

o

Figure 1. EffectsofDrugson ~APPCarboxy-TerminalFragmentsand on 13-Amyloid Peptides BC1 antiserum was used to immunoprecipitate I~APP carboxy-terminal fragments from cell lysates (A), and BA2 antiserum was used to immunoprecipitate I~-amyloid peptides present in media (B). Lane 1, no drug treatment; lane 2, 200 I~M MDL 28170 inhibitor; lane 3, 200 I~M leupeptin; lane 4, 200 I~M E64; lane 5, 100 p.M chloroquine; lane 6, 30 mM NH4CI; lane 7, 30 mM NH,-acetate; lane 8, 30 mM methylamine. Molecular weight markers (M) are indicated. The four major 13APP carboxy-terminal fragments, ranging in molecular weight from 8.0 to 12.0 kDa, that accumulate from MDL inhibitor treatment are identified in (A), and the 4 and 3 kDa 13-amyloid peptides are indicated in (B). MDL 28170 structure is in ((3). Treatment with each drug was for 4 hr and was simultaneous with [~S]methionine/cysteine labeling.

(CHO) cells stably transfected with the 695 amino acid isoform of 13APP as a source of 13-amyloid peptides and a system for study. A clonal line from the CHO cell transfec- tion (CP-6) was metabolically labeled with [35S]methionine/ cysteine for 4 hr in the absence and presence of various drugs. The set of compounds applied included: inhibitors of lysosomal cysteine proteinases, leupeptin and E64; acidotrophic agents that neutralize lysosomes and other acidic compartments such as endosomes and late and medial Golgi compartments, chloroquine, NH4-acetate, NH,CI, and methylamine; and the calpain proteinase inhibitor, MDL 28170 (Mehdi et al., 1988; Mehdi, 1991). The chemical structure of the MDL inhibitor, an amino- terminal blocked dipeptide aldehyde, is displayed in Fig- ure 1C. After the labeling period, cells and conditioned media were harvested. Cell iysates were prepared, I~APP carboxy-terminal fragments were immunoprecipitated with an antiserum directed to the cytoplasmic domain of ~APP

(BC1), and the precipitates were analyzed by polyacryi- amide gel electrophoresis using a Tris-Tricine buffer system. Conditioned media were immunoprecipitated with an antiserum raised to synthetic 13-amyloid 1-40 peptide, BA2. This latter antiserum has been fully characterized with respect to its specificity for 13-amyloid and the epitope recognized within 13-amyloid. Synthetic ~-amyloid peptide representing residues 1-40 can successfully compete the immunoprecipitation of [35S]methionine-labeled 4 and 3 kDa I~-amyloid peptides released from cultu red cells (data not shown). The 13-amyloid peptides were analyzed by 16.5% Tris-Tricine polyacrylamide gel electrophoresis.

The effect of each compound on 13APP carboxy-terminal fragment and p-amyloid peptide production is shown in Figure 1. Leupeptin and E64 had no effect on the levels of carboxy-terminal fragments generated during the 4 hr labeling period, and neither drug reduced levels of 4 and 3 kDa I~-amyloid peptides as compared with the untreated culture. Treatment with NH4CI, NH4-acetate, and methylamine each resulted in an increase in an 8 kDa 13APP carboxy-terminal fragment, as well as a modest reduction of I~-amyioid peptides. Chloroquine gave similar results to NH,CI, NH4-acetate, and methylamine. In contrast, the MDL inhibitor completely inhibited the formation of both 4 and 3 kDa I~-amyloid peptides. In addition, the MDL inhibitor caused an accumulation of four prominent 13APP carboxy-terminal fragments ranging in molecular weights from 8.0 to 12.0 kDa. Several minor larger fragments were also enhanced by treatment with the MDL inhibitor.

Characterization of pAPP Carboxy-Terminal Fragments The four predominant J3APP carboxyl termini that accumulated as a result of inhibition of I~APP processing by the MDL inhibitor were each characterized by radiosequencing. CP-6 cells were metabolically labeled with either [3H]alanine or [3H]phenylalanine in the presence of the M DL 28170 inhibitor, after which these 8.0-12.0 kDa 13APP carboxy-terminal fragments were separated by electrophoresis on 16.5% Tris-Tricine polyacrylamide gels, isolated, and microsequenced. The fastest migrating fragment, with an apparent molecular weight of 8.0 kDa, has an amino terminus beginning with leucine-17 of the 15-amyioid peptide. We observed [3H]alanine in cycles 5 and 14, and [3H]phenylalanine in cycles 3 and 4 (Figure 2A). The next fastest migrating 13APP fragment of 8.8 kDa was found to initiate with tyrosine-10 of 13-amyloid, since the radio-amino acid sequencing yielded [3H]phenylalanine at cycles 9 and 10 and [~H]alanine at cycles 11 and 20, as well as at cycles 5 and 14 (Figure 2B). The [3H]alanine seen in cycles 5 and 14 is likely to be due to contamina- tion from the very abundant adjacent 8.0 kDa fragment. Sequencing of the third and 10.0 kDa ~APP carboxy- terminal fragment revealed that this fragment possesses the mature amino terminus of 13-amyloid commencing with aspartate-l. Upon microsequencing of this fragment, [3H]phenylalanine and [3H]alanine were observed in cycles 4 and 2, respectively (Figure 2C). The largest I~APP carboxy-terminal fragment of the set was calculated at an apparent molecular weight of 12.0 kDa. Sequencing of

J~-Amyloid Processing 653

A

B

C

D

1600- 1400 Z

~ 1 2 0 0

~,~1000- .~ 600-

~ - 600-

400- 200-

O

2~! 114 0 ~ 100

~ 80 ~ 60 ~ 40

2000-

~ 1500

1000 -

~:~ 500-

5 15 20 25

4°°1

120-

100-

80-

,~ 60-

~" 40- ~n

20-

5 10 15 20 25 30 Cycle Number

20 NH IIIIIIHIIHF II I 0 • 5 10 15 20

8

1 5 10 15 20 25

~4

7C

6O

4O

-~ 30 20

1{3

1 2 5 4 5 6 7 5 Cycle Number

E ®

. . . LT~IKTEE I SEVKM~AE .......

12.0 kDa 10.0 kDa

...... FRHD S GTYEVHHQK~VFFA.. "

8,8 kDa 8.0 kDa

Figure 2. Radiosequence Analysis of ~SAPP Carboxy-Terminal Fragments Accumulating from Inhibition of ~-Amyloid Processing by the MDL 28170 Inhibitor (A) Analysis of the 8.0 kDa fragment using [3H]phenylalanine and [3H]alanine. (B) Analysis of the 8.8 kDa fragment using [3H]phenylalanine and [SH]analine. Asterisks indicate contaminating radioactive peaks from the 8.0 kDa fragment. (C) Analysis of the 10.0 kDa fragment using [3H]phenylanaline and [SH]alanine. (D) Analysis of the 12.0 kDa fragment using pH]phenylanaline. (E) Partial i~APP amino acid sequence illustrat- ing cleavage sites yielding the 8.0, 8.8, 10.0, and 12.0 kDa fragments (arrows). The first residue of the 4 kDa ~-amyloid is also indicated (1).

this 12.0 kDa fragment gave a peak of [3H]phenylalanine at cycle 16, indicating that the amino terminus of this fragment is extended by 12 residues from the mature 13-amyloid amino terminus (or at asparagine-585 of the ~APP(395 isoform sequence). Sequencing results for the 12.0 kDa fragment labeled with [3H]alanine were inconclu- sive owing to the low yield of this proteolytic product. In support of these data, when this same set of four 6APP carboxyl termini was prepared unlabeled (employing a different ~APP expression system and cell background), and when the amino-terminal sequence of each fragment was determined directly, the sequences obtained were in agreement with these results (data not shown). Both the 10.0 and 12.0 kDa fragments qualify as potential precursors to the 4 kDa ~-amyloid peptide.

Identification of Proteolytic Precursors to I~-Amyloid Peptides To examine the possible precursor-product relationship between these ~APP carboxy-terminal fragments and the ~-amyloid peptides, we took advantage of the fact that the effects of the MDL 28170 inhibitor can be reversed by simply washing the cell monolayers. CP-6 cell cultures were treated with the MDL inhibitor for 1 hr, during which time they were also metabolically !abeled with [3sS]methionine/cysteine. After the inhibitor treatment and labeling

period, the cells were washed and placed in medium lacking radioactivity and inhibitor. Cells and su pernatants were harvested at 0, 30, 60, 120, and 180 min after removal of the inhibitor and isotope, then analyzed for 13APP carboxy- terminal fragments and 13-amyloid peptides. The results from this inhibitor wash-out experiment are presented in Figure 3. At the commencement of the chase period, accumulation of the 8.0-12.0 kDa 13APP carboxyl termini and inhibition of 13-amyloid peptides were evident. However, 30 min after removal of the MDL inhibitor, there was a loss of carboxy-terminal fragments with a concomitant formation of the 4 and 3 kDa 13-amyloid peptides. By 120 min after removal of inhibitor, only trace amounts of carboxyl termini remained, and levels of I~-amyloid peptide were nearly equivalent to that produced by control cells which were not treated with the MDL inhibitor but were pulse labeled for 1 hr and then chased for 180 min. At 180 min, control- and M DL inhibitor-treated samples were not identical in that processing of drug-treated carboxy-terminal fragments occurred faster. The reason for this temporal discrepancy is not clear.

Degradation of p-Amyloid Proteolytic Precursors We were interested in the fate of the 13APP carboxy- terminal fragments that accumulate upon MDL inhibition of 13-amyloid peptide processing, in particular, whether

Neuron 654

A kDa M CON 0 30 60 120 180rain.

29

A MDL

1 r

+

2 L 7 r - + -

4 6 8 7 F I I 7 F + - + - + -

10hr . I

+ M kDa

18

14

B kDa M CON 0 30 60 120 180min. 18

14

Figure 3. Demonstration of Precursor-Product Relationship between [BAPP Carboxy-Terminal Fragments and I)-Amyloid Peptides using MDL Inhibitor Wash-Out (A) [BAPP carboxy-terminal fragments immunoprecipitated with BC1 antiserum from cell lysates; (B) J~-Amyloid peptides immunoprecipitated from conditioned media with BA2 antiserum. Cellular protein was metabolically labeled with [~S]methionine/cysteine for 60 rain in the presence of 200 I~M MDL 28170 inhibitor. After labeling, cells were washed free of label and inhibitor and then chased in isotope- and inhibitor-free medium. Samples were harvested at the chase times indicated (0-180 min), immunoprecipitated, and analyzed as described in the text. A sample (CON) that was labeled for 60 min and chased for 180 min in the absence of the MDL 28170 inhibitor served as a control. Molecular weight standards (M) are indicated.

these f ragments can be al ternat ively degraded in the con- t inued presence of the MDL 28170 inhibitor. To examine this possibil i ty, we employed two exper imenta l paradigms; first, a radiolabel pulse for protein synthesis fo l lowed by a nonradioact ive chase period to fol low protein turnover, and second, cont inuous protein radio label ing to observe steady-state protein levels. For the radiolabel pulse-chase exper iments, CP-6 cells were pulse labeled with [3~S]methionine/cysteine for 1 hr in the presence or absence of the MDL inhibitor, after which the cells were placed in medium lacking radiolabel, again in either the presence or absence of the MDL inhibitor. Samples were harvested at 1, 2, 4, 6, 8, and 10 hr into the chase. Cell lysates were immunoprecipi tated with BC1 ant iserum to character ize I~APP carboxy-terminal f ragments; med ia were immunoprec ip i ta ted with BA2 ant iserum to assay 13-amyloid peptides. As ex- pected, in the presence of the MDL inhibitor, carboxy- terminal f ragments dramat ical ly accumulated ( - 50 - f o l d over levels in untreated cultures) during the first 2 hr of the

B MDL F_

I r

2 1 r ÷ -

4

+

6 8 7 r I r + - + -

10 hr. 1 + M kDa

C 1 4 8 24 hr. F - - I F - - 7 I - - 3 F - -

MDL - + + - + - + I

M kDa

29

18

14

D 4000-

3000 -

22000 -

1000 -

0

0 - #" ~ L

5 10 ll5 20 25 Time (hours)

Figure 4. Effect of MDL Inhibition of 13APP Carboxy-Terminal Frag- ments and 13-Amyloid Peptides in a Pulse-Chase Experimental Format (A and C) Accumulation and degradation of pulse-labeled I)APP carboxy-terminal fragments over 10 hr (A) and 24 hr (C) chase periods. ([A] was intentionally overexposed to emphasize minor degradation fragments.) (B) Inhibition of pulse-labeled [B-amyloid peptides by the MDL 28170 inhibitor over a 10 hr chase period. The (+) and (-) denote the presence and absence of 200 I~M MDL 28170 inhibitor, respectively. Cellular protein was metabolically labeled with [~S]methionine/cysteine for 1 hr, after which isotope was removed and cells were placed in isotope- free medium (time 0 of the chase period). A graphical representation of the data in (C) is presented in (D). Open and closed circles indicate addition or absence, respectively, of the MDL 28170 inhibitor.

13-Amyloid Processing 655

0 r - - [ F - -

m MDL +

2

,t-

6 8 l r - - P F - -

- + +

24 hr. l - -

+ ]

M kDa

29

18

14

0 2 f - - I F - -

B MDL +

6 8

+ - ~ - +

24 hr. q

+ M kDa

18

14

Figure 5. Effect of the MDL 28170 Inqibitor on ~APP Carboxy-Ter- minal Fragments and ~-Amyloid Peptides in a Continuous Labeling Experimental Format I3APP carboxy-terminal fragments (A) and J3-amyloid peptides (B) produced upon continuous labeling with pS]methionine/cysteine in either the absence (-) or the presence (+) of 200 ~M MDL inhibitor. Samples were harvested at the times indicated over a 24 hr labeling period.

chase period, after which time the levels of the fragments decreased. By 10 hr, an 8-fold reduction from initial accumulated levels could be determined (Figure 4A). New, smaller (5-7.5 kDa) immunoprecipitable fragments ap- peared at4 hr into the chase period and were present at all later time points. As the J3APP carboxy-terminal fragments were gradually depleted over the chase period, complete inhibition of 13-amyloid peptide formation was sustained throughout the entire 10 hr chase period, indicating that the degradation of MDL inhibitor-induced carboxy-terminal fragment accumulation was to nonamyloidogenic products (Figure 4B). When this experiment was repeated and the fate of 13APP carboxy-terminal fragments was followed out to 24 hr, only a small residual amount (13%) of 13APP fragments remained after 24 hr (Figures 4C and 4D). Con- tinuous labeling experiments carried out in the presence or absence of the MDL inhibitor revealed a nearly complete inhibition of ~-amyloid peptides at all time points (Figure 5B). Also, continuous labeling after 2 hr produced a com- posite of immunoprecipitated proteins reflecting the accumulation of amyloidogenic carboxyl termini, as well as the novel 5-7.5 kDa catabolic fragments noted above in the pulse-chase fragment turnover experiments (Figure 5A). When these 5-7.5 kDa J~APP carboxy-terminal derivatives were analyzed by immunoprecipitation with several ~-amyloid peptide antisera, including BA2, none reacted,

indicating that they were degraded to non-l~-amyloid peptide products.

Because the 13APP proteolytic intermediates that accumulated due to inhibition of 13-amyloid processing by the MDL 28170 inhibitor were eventually degraded to nonamy- Ioidogenic products, we determined the effect of various drugs capable of inhibiting different subcellular compartments, to identify the locale where these fragments are metabolized. All of the following experiments were identical in design. A pulse-chase format was followed in which CP-6 cell cultures were radiolabeled with [3~S]methionine/ cysteine for 1 hr, after which label was removed and the cultures were incubated for chase periods of 0, 4, 8, 12, and 24 hr. The MDL inhibitor was present in all cultures during the entire pulse-chase period. Multiple pairs of CP-6 cell cultures were prepared in this fashion. To one set of cultures, a second inhibitor was added at the initiation of the chase period, in addition to the MDL 28170 inhibitor; the other set was treated with only MDL inhibitor. The inhibitors assayed in this experiment included: two inhibitors of lysosomal proteinases, leupeptin and E64; two acidotropic agents, chloroquine and NH4CI, which neutralize lysosomes and other acidic compartments such as endosomes and the trans-Golgi network; monensin, a car- boxylic ionophore that blocks protein processing and transport within the distal Golgi and that also disrupts endosome/lysosome function; and brefeldin A, which blocks protein transport by causing the resorption of the Golgi into the endoplasmic reticutum.

In this series of experiments, we found that the thiol proteinase inhibitors, leupeptin and E64 (Figures 6B and 6C), had no effect on MDL inhibitor-induced I3APP carboxy-terminal fragment catabolism compared with control cultures (Figures 6A and 6E). Similarly, brefeldin A did not prevent the degradation of these fragments (Figure 6D). In contrast, chloroquine and NH4CI both retarded the clear- ance of ~APP termini (Figures 6F and 6G), while monensin prevented the degradation of these fragments (Figure 6H). The catabolism of the 10.0 and 12.0 kDa amyloidogenic ~APP fragments was inhibited by monensin, and these fragments persisted throughout the 24 hr chase period. In a different experiment, accumulated pulse-labeled carboxy-terminal precursors, when washed free of the MDL inhibitor, were not processed into J3-amyloid peptides when monensin was present (data not shown), suggesting that the compartments involved in processing the carboxyl termini of ~-amyloid peptides and the degradation of I~APP fragments are the same. We also utilized reduced temperatures to halt 6APP processing at the endoplasmic reticulum (15°C) and at the trans-Golgi network (23°C) in the presence and absence of the MDL inhibitor, in the presence of inhibitor, carboxy-terminal fragment accumulation occurred only at 37°C. The formation of ~-amyloid peptides was seen only in the absence of inhibitor at 37°C and not at the lower temperatures (data not shown). These data, together with those derived from chemical inhibition of protein transport, indicate that a post-Golgi, pre-lysosomal compartment is involved in "y-secretase processing of 13APP carboxy-terminal fragments to ~-amyloid.

Neuron 656

A 0 4 8 12 24 B CON LEU

C 0 4 8 12 24 D E64 BFA

E 0 4 8 12 24 F CON CLQ

0 4 8 12 24 hrs

0 4 8 12 24

0 4 8 12 24

Figure 6. Effects of Drugs on the Degradation of 13APP Carboxy-Terminal Fragments Accumu- lating from Inhibition of ~-Amyloid Processing by the MDL Inhibitor Cells were pulse labeled with [3SS]methionine/ cysteine for 1 hr in the presence of 200 I~M MDL 28170 inhibitor. At the start of the 24 hr chase period, a second inhibitor was added tothe media containing the MDL 28170 inhibitor. Samples were taken for analysis at the times shown (0, 4, 8, 12, and 24 hr). (A) MDL inhibitor only (CON); (B) MDL inhibitor plus 200 p.M leupeptin (LEU); (C) MDL inhibitor plus 200 p.M E64 (E64); (D) MDL inhibitor plus 10 p.M brefeldin A (BFA); (E) MDL inhibitor only (CON); (F) MDL inhibitor plus 100 #.M chloroquine (CLQ); (G) MDL inhibitor plus 30 mM NH4CI (NH,CI); (H) MDL inhibitor plus 10 p.M monensin (MON). All control samples were approximately identical; onlytwo are shown for space purposes, The migration of internal molecular weight standards is indicated.

G 0 4 8 12 24 H

NH4CI MON

0 4 8 12 24

i

Discuss ion

The finding that 13-amyloid peptide is produced by normal physiological processes and is released by cultured cells (Shoji et al., 1992; Haass et al., 1992b; Busciglio et al., 1993) has provided a system to enhance our understand- ing of the molecular mechanisms leading to the formation of this 4 kDa derivative of 15APP that is deposited in AIzhei- mer's disease. The location of the I~-amyloid sequence within I~APP dictates the involvement of proteolytic cleavage events in the formation of this peptide. Two pathways have been proposed as potential candidates for 13-amyloid production, the secretory and lysosomal pathways. Amy- Ioidogenic fragments of I~APP have been identified as proteolytic processing intermediates of both pathways (Seu- bert et al., 1993; Haass et al., 1992a; Estus et al., 1992; Golde et al., 1992; Siman et al., 1993). However, it has not been directly demonstrated that any naturally occurring I~APP amyloidogenic processing products are authentic precursors of I~-amyloid peptide. Results from pharmacological studies suggest that 13-amyloid peptide is gener-

ated within the cellular secretory pathway, since the genesis of the peptide is sensitive to agents that block protein transport through the secretory pathway (Busciglio et al., 1993; Haass et al., 1993). However, it does not appear that the proteolytic cleavage that produces secreted 13APP is involved in I~-amyloidogenesis. Agents that stimulate protein kinase C and enhance secretion of 13APP, such as phorbol esters, reduce levels of the 4 kDa 13-amyloid peptide while simultaneously increasing levels of the 3 kDa amyloid peptide 13APP secretion by-product (Bux- baum et al., 1993; Hung et al., 1993; Gabuzda et al., 1993). Presumably, for a given amount of I~APP substrate, en- hancing I~APP secretion, and thus the 3 kDa peptide, de- pletes I~APP substrate available for 4 kDa I~-amyloid production.

Furthermore, lysosomal proteolysis is not responsible for the formation of the 13-amyloid peptides, since inhibitors of lysosomal proteinases, E64 and leupeptin, do not block their formation (Shoji et ai., 1992; Haass et al., 1993; Bus- ciglio et al., 1993; results herein). However, 13-amyloid peptide genesis is sensitive to acidotrophic agents, such as

J3-Arnyloid Processing 657

NH4CI and chloroquine, indicating that a weakly acidic subcellular compartment is involved (Shoji et al., 1992; Haass et al., 1993; Busciglio et al., 1993; results herein).

To define better the events leading to ~-amyloid peptide formation, we have applied an inhibitor, MDL 28170 (Mehdi, 1991), that blocks processing of ~APP to I~-amyloid peptides. Application of the MDL inhibitor to cultured mammalian cells was found to inhibit the production of both the 4 and 3 kDa ~-amyloid peptides, while concomitantly producing an accumulation of 13APP ca~'boxy-terminal fragments ranging in size from 8.0 to 12.0 kDa. The MDL 28170 inhibitor blocks the processing of these 6APP fragments by preventing the cleavage that produces the carboxyl termini of the 13-amyloid peptides. In other words, the MDL inhibitor appears to inhibit ¥-secretase activity. The MDL 28170 inhibitor as described by Mehdi (1988, 1991) is a calpain proteinase inhibitor; however, no direct evidence exists that ~,-secretase is a thiol proteinase, or is a calcium- activated thiol proteinase.

A direct precursor-product relationship between the 8.0-12.0 kDa carboxyl termini of 13APP and the 4 and 3 kDa I~-amyloid peptides was demonstrated. Removal of the MDL 28170 inhibitor from the cells resulted in the grad- ual disappearance of these 13APP fragments, with a recip- rocal production of both the 4 and 3 kDa ~-amyloid peptides. Additional supporting evidence for the precursor role of these 13APP fragments was obtained from sequence analysis. The predominant 8.0 kDa fragment was found to have an amino terminus consistent with the major cleavage site used for 13APP secretion, i.e., leucine-17 of the 13-amyloid peptide (Esch et al., 1990; Anderson et al., 1991 ; Zhong et al., 1994). Therefore, this 8.0 kDa carboxy- terminal ~APP fragment is the precursor to the 3 kDa peptide, since Busciglio et al. (1993) and Haass et al. (1992b) characterized the amino term in us of a major 3 kDa species as initiating with leucine-17. Other 3 kDa peptides with truncated amino termini have also been described (Bus- ciglio et al., 1993; Haass et al., "!992b). The amino terminus of a minor 8.8 kDa carboxy-terminal fragment that accumulates with MDL 28170 inhibitor treatment does not correspond to any of the reported 3 kDa peptide sequences. This fragment would produce a peptide bearing an amino terminus of tyrosine-10. The amino-terminal heterogeneity observed by sequence analysis of the 3 kDa secreted products may reflect heterogeneity in the proteolytic cleavage leading to ~APP secretion, as has been reported (Zhong et al., 1994), and/or might result from exoaminopeptidase activity or additional proteolytic processing following an initial set o'I cleavages.

The amino terminus determined for a 10.0 kDa carboxy- terminal precursor fragment corresponds to the mature amino terminus of J3-amyloid, aspartate-l. The amino terminus of this fragment is in agreement with that reported for the major 4 kDa amyloidogenic peptide produced in vitro and in vivo (Haass et al., 1992b; Shoji et al., 1992; Seubert et al., 1992). Additional 4 kDa species have also been described that reflect amino-terminal extensions (Busciglio et al., 1993) or truncations (Haass et al., 1993). The presence of 13-amyloid peptides with amino-terminal

extensions are compatible with the larger 12.0 kDa amy- Ioidogenic carboxy-terminal IBAPP fragment that accumulates when 13-amyloid processing is inhibited.

The coinhibition of the 4 and 3 kDa peptides by the MDL 28170 inhibitor suggests that a common cellular processing compartment is used for the carboxy-terminal processing that generates the two peptides. The most likely subcellular site(s) is the trans-Golgi network or early endosome. Experiments involving the inhibitor brefeldin A re- veal that transport of 13APP through the Golgi is a requisite for formation of the J3-amyloid peptides (Haass et al., 1993; Busciglio et al., 1993). Monensin, an ionophore that neu- tralizes acidic intracellular compartments and promotes the swelling of the Golgi, especially the trans-Golgi network, also inhibits J3-amyloid formation. Other acidotrophic agents, such as chloroquine and NH4CI, have partial to full inhibitory effects on the production of these peptides (Shoji et al., 1992; Haass et al., 1993; Busciglio et al., 1993), also indicating that a mildly acidic compartment such as the distal Golgi complex or early endosome is involved. These same cellular compartments have been shown to be sites of proteolytic cleavage leading to ~APP secretion (Kuentzel et al., 1993; Sambamurti et al., 1992; De Strooper et al., 1993). Hence, 3 kDa precursor fragments derived from 13APP secretion would be present in the same cellular compartment(s) that harbor the 4 kDa peptide precursor fragments. The trans-Golgi network and/or post-Golgi compartment have been identified as the primary site(s) of ~APP secretory cleavage (Kuentzel et al., 1993; Sambamurti et al., 1992; De Strooper et al., 1993). J3APP secretory cleavage on the cell surface, a mi- nor cleavage site (Kuentzel et al., 1993), would also bring carboxy-terminal remnants into an endosomal compartment through protein internalization. Thus, while both amyloidogenic and nonamyloidogenic carboxy-terminal 6APP fragments may reflect different initial processing events, it appears that these fragments are ultimately processed together to yield both the 3 and 4 kDa peptides.

We found that MDL 28170 inhibitor treatment results in the accumulation of ~APP carboxy-terminal precursors in a subcellular compartment that can degrade these fragments to other than ~-amyloid peptide products. This catabolism is insensitive to leupeptin and E64 treatment, precluding the lysosome as the responsible subcellular site. Similarly, brefeldin A did not perturb the degradation of the preamyloid fragments, thereby excluding the endoplasmic reticulum and the Klausner degradative pathway (Lippincott-Schwartz et al., 1988) as a potential site of catabolism. The acidotropic agents, chloroquine and NH4CI, retarded the degradation of these fragments, indicating that a weakly acidic compartment is involved. Monensin, which can neutralize acidic compartments but which also disrupts distal Golgi cisternae, completely blocked degradation. Monensin was also found to disrupt the compartment and/or proteinase activity responsible for producing the cleavage that forms the carboxyl termini of the 13-amyloid peptides. In addition, using reduced temperature to block intracellular ~APP transport indicated that a post-Golgi compartment is responsible for the carboxy-terminal pro-

Neuron 658

cess ing of J3-amyloid. The re fo re , the in te rp re ta t ion most

compa t i b l e with t hese data, as wel l as t h o s e de f in ing

~-amylo id pep t ide fo rmat ion , po in ts to an ear l y e n d o s o m e

as the p r o b a b l e site of 13APP ca tabo l i sm and 15-amyloid

pep t ide p roduc t ion . Last ly, it is impor tan t to u n d e r s c o r e the obse rva t i on that

sus ta i ned inh ib i t ion of 13-amyloid f o rma t ion leads to an al-

t e rna te m e a n s o f d e g r a d a t i o n o f a m y l o i d o g e n i c p recu rso r

f ragmen ts . Fu r the rmore , w e f ind that 13APP b iosyn thes is

and sec re t i on a re una f fec ted by t r ea tmen t with the M D L

inh ib i tor ( unpub l i shed data). T h e s e resul ts sugges t that

I~-amyloid is f o r m e d as a nonob l i ga to r y d e g r a d a t i v e by-

p roduc t f rom the ca tabo l i sm o f ma tu re 13APP. M o r e im-

por tant ly , t hese da ta have pos i t i ve imp l ica t ions for the po-

tent ia l t r ea tmen t of A lzhe imer ' s d isease .

Experimental Procedures

Cell Culture, Radiolabeling, and Drug Treatment The CP-6 cell line was established by transfecting the human cDNA encoding the 695 amino acid isoform of 13APP, driven by the human 13-actin promoter (Leavitt et al., 1984), into CHO cells using calcium phosphate according to standard procedures. Individual transfectants were isolated, subcloned, and characterized for expression of exoge- nous 13APP by Western blotting and immunoprecipitation procedures. CP-6 cells were propagated in a 1:1 mixture of Dulbecco's minimum essential medium and Coon's F12 medium with 10% fetal bovine serum in 10 cm dishes ( - 1 x 106 cells per dish). To metabolically radiolabel protein with [35S]methionine/cysteine (Translabel, iCN), CP-6 cell monolayers were washed three times with saline and placed in serum-, methionine-, and cysteine-free medium for 30 min, after which time 167 pCi/ml [~S]methionine/cysteine was added. Culture dishes were radiolabeled for various times depending on the experimental design. For pulse-chase experiments, cultures were labeled for 60 min, and cell monolayers were washed, placed in fresh media lacking isotope (chase), and incubated for varying lengths of time prior to harvesting. For both pulse-chase and continuous labeling experiments, it was determined that the low levels of methionine and cysteine were neither deleterious nor limiting to the cells during 24 hr. For radiosequencing experiments, CP-6 cells were incubated for 30 min in serum-free medium lacking the appropriate amino acid and then labeled with - 750 llCi/ml [3H]alanine (Amersham) or - 750 iiCi/ml [3H]phenylalanine (Amersham) in serum-free medium lacking alanine or phenylalanine, respectively, for 4 hr prior to harvesting.

For experiments involving treatment with drugs, the following con- centrations were used: 200 ~M MDL 28170 inhibitor (Mehdi et al., 1988; Mehdi, 1991), 200 I~M leupeptin (Boehringer Mannheim), 200 I~M E64 (Boehringer Mannheim), 100 ~M chloroquine (Sigma), 30 mM NH4CI, 30 mM NH4-acetate, 30 mM methylamine, 10 p_M monensin (Calbiochem), and 10 p_M brefeldin A (Epicentre Technologies, Madi- son, Wl). Time of application of each drug was variable depending on the specific experiment.

Inquiries regarding access to the MDL 28170 inhibitor should be made with the corresponding author.

Antibodies A polyclonal antibody to a region in the I~APP cytoplasmic domain (position #704-730 of the I~APPTsl sequence) was prepared by repeated immunization of rabbits with a conjugate of purified synthetic peptide and keyhole limpet hemocyanin. Polyclonal antisera to 13-amy- Ioid were raised in rabbits using a synthetic 13-amyloid 1-40 peptide purchased from Bachem (Torrance, CA). The peptide was conjugated to keyhole limpet hemocyanin. Both I~APP carboxy-terminal and 13-amyloid antibodies were characterized for reactivity against their respective cognate immunogen in an enzyme-linked immunoadsor- bent sandwich assay, by immunoprecipitation of radiolabeled !SAPP fragments or ~-amyloid, and for the ability of their cog nate immunogen to block immunoprecipitation of the specific ~APP product.

Sample Preparation, Immunoprecipitation, and Analysis Immunoprecipitation of conditioned medium (3 ml total) for ~-amyloid peptide detection was performed as follows. RIPA buffer (7 ml; 50 mM Tris-HCI [pH 7.5], 150 mM NaCI, 1% deoxycholate, 0.1% SDS, 1% Triton X-100) was added to clarified medium and precleared with 45 I~1 of normal rabbit serum plus 100 I~1 of protein A-Sepharose (Phar- macia) by rocking at 23°C for 2.5 hr or 4°C overnight followed by centrifugation at low speed. Protein A-Sepharose (100 p.I) and p-amy- Ioid antiserum, BA2 (45 Id), were added to the low speed supernatant and incubated as above. The mixture was again centrifuged at low speed, and the pellet was washed twice with 50 mM Tris-HCi (pH 7.5), 500 mM NaCI, 5 mM EDTA, 0.5% Nonidet P-40, twice with 50 mM Tris-HCI (pH 7.5), 150 mM NaCI, 5 mM EDTA, 0.5% Noniclet P-40, and twice with 10 mM Tris-HCI (pH 7.5). The entire pellet was resuspended in SDS reducing sample buffer and loaded onto one lane of a 16.5% Tris-Tricine SDS-polyacrylamide gel (Zhong et al., 1994). For immunoprecipitation of ~APP carboxy-terminal fragments, cell lysates were prepared according to the method of Gabuzda et al. (1991). Each cell lysate, diluted to 5 ml with RIPA buffer, was precleared with normal rabbit serum and protein A-Sepharose as described above. ~APP carboxy-terminal antiserum, BC1 (30 p_l), and protein A-Sepha- rose (100 ~1) were added to the lysate mixture, after which the immunoprecipitation and gel electrophoresis steps followed the same procedures as described above for conditioned medium samples. All gels were developed using autoradiography. Film exposure times were - 4-6 days for I~-amyloid peptides and overnight for 13APP carboxy- terminal fragments, Densitometry was carried out using a Phosphorlm- ager (Molecular Dynamics).

Amino Acid Sequence Analysis Carboxy-terminal fragments were prepared by immunoprecipitation and separated electrophoretically as described above. Protein fragments were transferred to polyvinylenedifluoride membranes, excised using [~S]methionine/cysteine-labeled reference ~APP fragments run on an adjacent lane of the gel, and subjected to Edman degradation using an Applied Biosystems A77A protein sequenator (Foster City, CA) using the BLOTT-2 program according to the manufacturer's rec- ommendations. Each degradation cycle was collected, mixed with scintillation fluid, and counted in a Beckman LS 3801 scintillation counter.

Acknowledgments

All correspondence should be addressed to B. C. We thank Norton Peet and Daniel Scherlin of Marion Merrell Dow, Inc., for providing the MDL inhibitor, Eric Stoelting for art work, and Linda Higgins for critical reading of this manuscript. We gratefully acknowledge Marion Merrell Dow, Inc., for sponsoring our research.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 USC Section 1734 solely to indicate this fact.

Received November 8, 1994; revised December 21, 1994.

References

Anderson, J. P., Esch, F. S, Kein, P. S., Sambamurti, K., Lieberburg, I., and Robakis, N. K. (1991). Exact cleavage site of Alzheimer amyloid precursor in neuronal PC-12 cells. Neurosci. Lett. 128, 126-128.

Busciglio, J., Gabuzda, D. H., and Yankner, B. A. (1993). Generation of 13-amyloid in the secretory pathway in neuronal and nonneuronal cells. Prec. Natl. Acad. Sci. USA 90, 2092-2096. Buxbaum, J. D., Koo, E. H., and Greengard, P. (1993). Protein phos- phorylation inhibits production of Alzheimer amyloid WA4 peptide. Prec. Natl. Acad. Sci. USA 90, 9195-9198. Caporaso, G. L., Gandy, S. E,, Buxbaurn, J. D., and Greengard, P. (1992). Chloroquine inhibits intracellular degradation but not secretion of Alzheirner ~/A4 amyloid precursor protein. Prec. Natl. Acad. Sci. USA 89, 2252-2256.

~-Amyloid Processing 659

Cataldo, A. M., and Nixon, R. A, (1990). Enzymatically active lysosomal proteases are associated with amyloid deposits in Alzheimer brain. Proc. Natl. Acad. Sci. USA 87, 3861-3665. De Strooper, B., Umans, L., Van Leuven, F., and Van Den Berghe, H. (1993). Study of the synthesis and secretion of normal and artifical mutants of murine amyloid precursor protein (APP): cleavage of APP occurs in a late compartment of the default secretion pathway. J. Cell Biol. 121,295-304. Esch, F. S,, Keim, P. S., Beattie, E. C., Blacher, R. W., Culwell, A. R., Oltersdorf, T., McClure, D., an:l Ward, P. J. (1990). Cleavage of amyloid 13 peptide during constitut ve processing of its precursor. Science 248, 1122-1124. Estus, S., Golde, T., Kunishita, T., Bl~,des, D., Lowery, D., Eisen, M., Usiak, M., Qu, X., Tabira, T., Greenberg, B. D., and Younkin, S. G. (1992). Potentially amyloidogenic, carboxyl-terminal derivatives of the amyioid precursor protein. Science 255, 726-728. Gabuzda, D. H., Olshevsky, U., Bertani, P., Haseltine, W. A., and Sodroski, J. (1991). Identification of membrane anchorage domains of the HIV-1 gp160 envelope glycoprotein precursor. J. AIDS 4, 34- 40. Gabuzda, D. H., Busciglio, J., and Yankner, B. A. (1993). Inhibition of p-amyloid production by activation of protein kinase-C. J. Neuro- chem. 61, 2326-2330. Golde, T. E,, Estus, S., Younkin, L, H., Selkoe, D. J., and Younkin, S. G. (1992). Processing of the amyloid precursor to potentially amy- Ioidogenic derivatives. Science 255, 728-730. Haass, C., Koo, E. H., Mellon, A., Hung, A, Y., and Selkoe, D. J. (1992a). Targeting of cell-surface 13-amyloid precursor proteins to lysosomes: alternative processing into amyloid-bearing fragments. Nature 357, 500-503. Haass, C., Schlossmacher, M. G., Hung, A. Y., Vigo-Pelfrey, C., Mel- lon, A., Ostaszewski, B. L., Lieberbu~'g, I., Koo, E. H., Schenk, D., Teplow, D. B., and Selkoe, D. J. (19~2b). Amyloid I~-peptide is produced by cultured cells during normal metabolism. Nature 359, 322- 325. Haass, C., Hung, A. Y., Schlossmacher, M. G., Teplow, D. B., and Selkoe, D. J. (1993). I~-Amyloid peptide and a 3-kDa fragment are derived by distinct cellular mechanisms. J. Biol. Chem. 268, 3021- 3024. Hayashi, Y., Kashiwagi, K., and Yoshikawa, K. (1992). Protease inhibitors generate cytotoxic fragments from Alzheimer amyloid protein precursor in cDNA-transfected glioma cells. Biochem. Biophys. Res. Commun. 187, 1249-1255. Hung, A. Y., Haass, C., Nitsch, R. M., Qiu, W. Q., Citron, M., Wurtman, R. J., Growdon, J. H., and Selkoe, D. J. (1993). Activation of protein kinase C inhibits cellular production of the amyloid 13-protein. J. Biol. Chem. 268, 22959-22962. Kuentzel, S. L., Ali, S. M., Greenberg, B. D., and Raub, T. J. (1993). The Alzheimer I~-amyloid protein precursor/protease nexin II is cleaved by secretase in a trans-Golgi secretory compartment in human neuro- glioma cells. Biochem. J. 295, 367-378. Leavitt, J., Gunning, P., Porreca, P., Ng S,-Y., Lin, C.-S., and Kedes, L. (1984). Molecular cloning and characterization of m utant and wild-type human J3-actin genes. Mol. Cell. Biol. 4, 1961-1969.

Lippincott-Schwartz, J., Bonifacino, J. S., Yuan, L. C., and Klausner, R. D. (1988). Degradation from the endoplasmic reticulum: disposing of newly synthesized proteins. Cell 54, 209-220.

Mehdi, S. (1991). Cell-penetrating inhibitors of calpain. Trends Bio- chem. Sci. 16, 150-153.

Mehdi, S., Angelastro, M. R., Wiseman, J. S., and Bey, P. (1988). Inhibition of the proteolysis of rat erythrocyte membrane proteins by a synthetic inhibitor of calpain. Biochem. Biophys. Res. Commun. 157, 1117-1123.

Sambamurti, K., Shioi, J., Anderson, J. P., Pappolla, M. A., and Ro- bakis, N. K. (1992). Evidence for intracellular cleavage of the Alzhei- mer's amyloid precursor in PC12 cells. J. Neurocsci. Res. 33, 319- 329.

Seubert, P., Vigo-Pelfrey, C., Esch, F., Lee, M., Dovey, H., Davis, D., Sinha, S., Schlossmacher, M., Whaley, J., Swindlehurst, C., McCor- mack, R., Wolfert, R., Selkoe, D., Lieberburg, I., and Schenk, D. (1992). Isolation and quantitation of soluble Alzheimer's 13-peptide from biologi- cal fluids. Nature 359, 325-327.

Seubert, P., Oltersdorf, T., Lee, M. G., Barbour, R., Blomquist, C., Davis, D. L., Bryant, K., Fritz, L. C., G alasko, D., Thai, L. J., Lieberburg, I., and Schenk, D. B. (1993). Secretion of J3-amyloid precursor protein cleaved at the amino terminus of the ~-amyloid peptide. Nature 361, 260-263. Shoji, M., Golde, T. E., Ghiso, J., Cheung, T. T., Estus, S., Shaffer, L. M., Cai, X.-D., McKay, D. M., Tinter, R., Frangione, B., and Younkin, S. G. (1992). Production of the Alzheimer amyloid 13 protein by normal proteolytic processing. Science 258, 126-129. Siman, R., Mistretta, S., Durkin, J. T., Savage, M. J., Loh, T., Trusko, S., and Scott, R. W. (1993). Processing of the 13-amyloid precursor. J. Biol. Chem. 268, 16602-16609. Sisodia, S. S. (1992). ~-Amyloid precursor protein cleavage by a membrane-bound protease. Proc. Natl. Acad. Sci. USA 89, 6075-6079. Zhong, Z., Higaki, J., Murakami, K., Wang, Y., Catalano, R., Quon, D., and Cordell, B. (1994). Secretion of I~-amyloid precursor protein involves multiple cleavage sites. J. Biol. Chem. 269, 627-632.

inhibition of-amyloid formation identifies proteolytic precursors and

Documents