THE JOURNAL OF BIOLCGICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 269, No. 16, Issue of April 22, pp. 12152-12158, 1994 Printed in U.S.A.

Modulation of AP Adhesiveness and Secretase Site Cleavage by Zinc*

(Received for publication, October 5 , 1993, and in revised form, December 22, 1993)

Ashley I. Bush, Warren H. Pettingell, Jr., Marc d. Paradis, and Rudolph E. TanziS From the Laboratory of Genetics and Aging, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts 02129

~-

Abnormalities of zinc homeostasis occur in Alzhei- mer’s disease (AD), a dementia characterized by the ag- gregation of AP in the brain, and in Down syndrome, a condition characterized by premature AD. We studied the binding of Zn2+ to a synthetic peptide representing residues 1-40 (AD,,), as well as other domains of AD. Two classes of Zn2+ binding were identified by @Zn2+ la- beling: highly specific pH-dependent high affinity (K, = 107 nM) binding, and lower affinity (K,, = 5.2 p ~ ) binding. Gel filtration chromatography identified monomeric, di- meric, and polymeric AP species. Zinc induced a marked loss of AP solubility upon chromatographic analysis. This was attributed to precipitation onto the column glass, which contains aluminosilicate, and was con- firmed by the observation of zinc-accelerated precipita- tion of AP by kaolin, a hydrated aluminum silicate sus- pension. Zinc binding also increased AP resistance to tryptic cleavage at the secretase site, indicating that a small (e3 p ~ ) increase in brain Zn2+ concentration could significantly alter AP metabolism. We propose that el- evated brain interstitial zinc levels may increase AP ad- hesiveness and interfere with AP catabolism. Conse- quently, abnormalities of regional zinc concentrations in the brains of patients with AD or Down syndrome may contribute to AP amyloidosis in these disorders.

Ab, a 4.3-kDa peptide, is the principal constituent of the cerebral amyloid deposits, a pathological hallmark of Alzhei- mer’s disease (AD)’ (Masters et al., 1985b; Glenner and Wong, 1984). A@ is derived from the much larger amyloid protein precursor (APP) (Kang et al., 1987; Tanzi et al., 1987; Robakis et al., 1987; Goldgaber et al., 19871, whose physiological func- tion remains unclear. The cause ofAlzheimer’s disease remains elusive; however, the discovery of mutations of APP close to or within the AP domain (Goate et al., 1991; Levy et al., 1990; Murre11 et al., 1991; Hendricks et al., 1992) indicates that the metabolism of AP and APP is likely to be intimately involved with the pathophysiology of this disorder.

* This work was supported by funds from the National Institutes of Health, the American Health Assistance Foundation, a Harkness Fel- lowship, Commonwealth Fund of New York (to A. I. B.), and a French Foundation Fellowship (to R. E. T.) 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 U.S.C. Section 1734 solely to indicate this fact.

netics and Aging, Massachusetts General Hospital, Harvard Medical $ To whom correspondence should be addressed: Laboratory of Ge-

School, Bldg. 149, 13th St., Boston, MA 02129. Tel.: 617-726-5746; Fax:

protein precursor; MOPS, 4-morpholinepropanesulfonic acid; MES, The abbreviations used are: AD, Alzheimer’s disease; APP, amyloid

4-morpholineethanesulfonic acid; BSA, bovine serum albumin; Tricine, N-[2-hydroxy-l,l-bis(hydroxymethyl)ethyllglycine.

617-726-5736.

Soluble AP is secreted in cell cultures and is found as a 40-residue peptide (AP,,,) in the cerebrospinal fluid (Shoji et al., 1992; Seubert et al., 1992; Haass et al., 1992). Physiological factors which can induce the aggregation of soluble AB are of interest in determining the cause of AP amyloid formation. Synthetic AP,,, remains soluble at concentrations up to 16 mg/ml in neutral phosphate buffer (Tomski and Murphy, 19921, indicating that overproduction of soluble AP does not SUE- ciently explain AP precipitation. Furthermore, soluble AP in cerebrospinal fluid is not increased in AD cases (Shoji et al., 1992), indicating that other pathogenetic mechanisms are likely to be involved.

Since one Zn2+ binding site in the APP ectodomain has al- ready been described (Bush et al., 1993), we investigated the possibility of additional zinc binding sites on APP. The AP,,,, structure possesses 3 histidines and several negatively charged residues, structural features that support Zn2+ binding. Here, we present studies showing that AP binds zinc in a saturable and specific manner. Moreover, we find that physiological con- centrations of Zn2+ increase the resistance of the peptide to proteolytic catabolism and promote AP precipitation by alumi- nosilicate. Based on these findings we propose tha t excessive zinc concentrations may serve to accelerate AP deposition in AD.

EXPERIMENTAL PROCEDURES Reagents-Precautions taken to avoid zinc contamination included

using analytical-grade reagents, electrophoresis-grade Tris-HC1 (Bio- Rad), and highly deionized water. A&,, was synthesized by the Biopolymers Laboratory, MIT. AP,,, (reverse peptide) was purchased from Bachem (Torrance, CA). Other reagents were from Sigma. A&,, and Ap,_,, results were replicated with peptides from Bachem and

W. M. Keck Foundation Biotechnology Resource Laboratory, Yale Uni- Sigma. Ap,,, results were also replicable with peptide synthesized by

versity. 65Zn was purchased from Amersham Corp. 65ZnZ+ Binding Studies-Dissolved peptides (1.2 nmol, unless other-

wise stated) were dot-blotted onto 0.2-pm polyvinylidene difluoride membrane (Pierce Chemical Co.), washed twice with chelating buffer (200 pl x 100 mM NaCl, 20 mM Tris-HC1,l mM EDTA, pH 7.41, then five times with blocking buffer (200 p1 x 100 mM NaCl, 20 II~M Tris-HC1,l m MnCl,, pH 7.4), and then incubated (60 min, 20 “C) with 65Zn (unless otherwise stated 130,000 cpm, 74 nM 65ZnC1, in 200 pl of blocking buffer

competing metal ion chloride). The dot-blot was then washed with blocking buffer (5 x 200 pl), the dot excised, placed in a test tube, and assayed by y-counting (11% efficiency). The equilibration volume for stoichiometry estimates was regarded as 6 x 200 pl. The 214 nm UV absorbance of the unbound flow-through was assayed to determine the total amount of peptide remaining bound onto the membrane. Peptide stock concentrations were confirmed by amino acid analysis. To alter the pH, the “Zn incubation was carried out in the presence of 100 mM buffer: MOPS (pH 6.5-7.01, MES (pH 5.0-6.0), acetate (pH 3.5-4.5). The dot-blot apparatus was washed with detergent and EDTA (50 m) then rinsed and siliconized between use.

Ap Chromatography-Ap (55 pg) was incubated with metal salt so- lution or EDTA in siliconized 1.5-ml plastic reaction vessels in 100 mM NaCl, 20 mM Tris-HC1, pH 7.4 (“TBS,” 100 pl, 1 h, 37 “C). AP was stored in aliquots of 0.52 mg/ml in water at -20 “C, then kept at 4 “C when

12152

Precipitation of AP by Zinc 12153

thawed. Reagents were mixed without vortex mixing. The incubated Ap was directly applied to a G50 SF (Pharmacia, Uppsala, Sweden) column (Bio-Rad Econo-Column, 30 x 0.7 cm) pre-equilibrated with metal salt solution or EDTA (50 PM) in TBS a t 20 "C and eluted a t 8 mL% (Wiz peristaltic pump, Isco, Lincoln, NE). Absorbance was measured a t 254 and 214 nm (Type 6 optical unit, Isco). The amount of AB eluting a t various peaks was estimated from the area under the curve. This was possible because the relationship of W absorbance was determined to be linear over the range ofAP dilutions used in these studies, indicating that absorbance is proportional to the amount of peptide present despite polymerization state (see below). The maximum recovery ofAp occurs in the presence of EDTA. Because the sample eluted in a volume of ap- proximately 15 ml, the average concentration of the peptide on the column was 0.8 PM.

To study the effects of protein blocking upon adsorption of Ap to the chromatography column, a Sephadex G50 SF column which had been characterized previously for A@ behavior was eluted with 3% bovine serum albumin (BSA) in TBS (50 ml) and equilibrated with non-BSA- ~ n t a i n i n g buffer, subsequent to repeating the Ap experiments.

Spectroscopic Assu~-Measurements were performed an a Hewlett- Packard 8452A diode array spectrophotometer using a 1-cm path length quartz cuvette. Concentration versus absorbance curves were per- formed a t 214 nm, 254 nm, 280 nm, and full spectrum. 214 nm readings were 50-fold more sensitive in detecting the peptide than 254 nm read- ings, whereas the 280 nm readings of low micromolar AD solutions were below sensitivity limits and hence could not be used in these studies. The standard curves generated were linear at concentrations below 0.1 mg/ml. In addition, the effects of Cu", Zn", EDTA. and TBS upon absorbance were examined. At concentrations below 0.1 mg/ml, adjust- ing the peptide in water to TBS caused - 15% quenching. Cu"-, Zn2+-, and EDTA-containing AB solutions were studied for artifactual absor- bance over the linear range of the 214 nm absorbance curve. 1 mM EDTA caused 60% quenching, hence 50 p~ EDTA was employed, contributing a similar degree of quenching to that observed with Cu2+ and Zn".

A@ Binding to Kaolin ~ A l ~ ~ i n ~ r n Silicute&"ofin suspension was prepared in high performance liquid chromatography water (Fisher), defined, and adjusted to 50% (v/v). Ap (40 pg) was incubated in sili- conized reaction vessels with either kaolin or Sephadex G50 SF (10 pl x 50% (v/v)) in CU", Zn", or EDTA (100 p1 in TBS, 5 min, room tem- perature). The suspension was then pelleted (1500 x g, 3 min) and the supernatant removed and diluted 20-fold with water to bring the W absorbance readings into the linear range. Samples were assayed a t 214 nm before and after incubation with kaolin or Sephadex.

Tlyptic Digestion ofA@Ap,,, (13.9 pg) was incubated with Zn2+ (12 pl x in blocking buffer, 1 h, 37 "C) and then digested with trypsin (12 ng, 3 h, 37 "C). The reaction was stopped by adding SDS sample buffer containing phenylmethylsulfonyl fluoride (1 m ~ ) , boiling the samples (5 mins), and applying the samples to TrisA'ricine gel electrophoresis and transfer. The blot was washed with EDTA, C~massie-stained, incu- bated with %Zn2+, ind~vidual bands were excised, assayed for "Zn2+ binding, and N-terminal sequenced to confirm the identity of the diges- tion products. The effects of ZnZ+ (up to 100 VM in TBS) on the activity of trypsin, itself, were assayed by assay of 2-Arg-amido-4-methylcou- marin (Sigma) fluorescent cleavage product and determined to be neg- ligible. 200 p Zn", however, inhibited tryptic activity by 12%.

RESULTS

To determine whether AP binds zinc, a synthetic peptide representing secreted A&,,, was incubated with @Zn2+. Rapid binding (60% B,,, at 1 min), which plateaued at 1 h (data not shown), was observed. Scatchard analysis of 65ZnZ+ binding de- scribes two saturable binding curves, a high affinity curve (Ka < 107 nM), and a lower affinity curve (Pia < 5.2 p ~ ) (Fig. la). Our affinity constant estimates might be skewed by assuming that the Tris buffer does not bind zinc. In fact, Tris-HC1 binds zinc and copper with stability constants of 4.0 and 2.6, respectively (Dawson et al., 1986). Incubating AP in the presence of higher concentrations of Tris (150 and 500 mM) abolishes 'j5ZnZ+ bind- ing to AP (-50% and -95%, respectively, data not shown), indicating that Tris-induced Zn2+ chelation cannot be excluded. Our calculated affinity constants are therefore upper limit es- timates.

"Zn2+ binding is very specific, with Znz+ being the only un- labeled metal ion tested that is capable of competing off the label (Fig. lbl . To determine the specific region of A@ involved

in zinc binding and to validate the dot-blot binding system, equivalent amounts of various peptides representing frag- ments of AP,,,, and peptide controls were assayed for 'j5Zn2' binding in this system (Figs. 1, c and d).

The reverse sequence (40-1) control peptide only binds 50% of B,,, compared with A&,, (Fig. IC), indicating that zinc binding is not merely a consequence of the presence of favorable residues. A@,_,, bound 30% ofB,,,, indicating that the carboxyl terminus plays an important role in promoting zinc binding. Glutamine substitution for the glutamate at position 11 of

in accordance with the Down syndrome AP sequence reported by Glenner and Wong (1984), does not interfere with G5Zn2+ binding. The Scatchard plot of 6eZn2" binding to AP,,, reveals similar lower affinity (KO < 15 J.IM) and higher affinity (Kc < 334 nM) binding associations (Fig. la! to those of A@,,,, but overall the Ap,-,, peptide binds zinc less avidly. Although the A&28 peptide clearly binds zinc, peptides overlapping this region (1-17 and 12-28) do not individually bind zinc. Addi- tionally, a peptide covering a region of the carboxyl terminus (25-35) also is unable to bind zinc (Fig. IC).

The calculated stoichiometry of high-affinity Zn2+-binding to AP, derived from the x-intercepts on the Scatchard plots (Fig. 1, a and d ) , is 0.7:l (API4J and 1:4 For low-affinity binding, the Zn":Ap ratio is 2 5 1 (Ap,,,) and 4:l

65Zn2+ binding of sequenced tryptic digest products ofAp (Fig. 4b) indicates that the 6 4 0 fragment binds zinc, but that the other visible digest fragment 17-40 (Fig. 4b ), representing the post-secretase (Esch et at., 1990; Sisodia et at: ~ 1990) carboxyl- terminal product, does not bind zinc. The contribution of histi- dines (residues 6,13, and 14) to En2' binding is indicated by the deterioration of binding with lower pH (30%, of Bmax a t pH 6.0, Fig. le). Taken together, these data indicate that zinc coordi- nation requires the contiguous sequence between residues 6 and 28, a region containing all 3 histidine residues, but optimal zinc binding also requires the presence of the carboxyl-terminal domain.

Next, we tested whether zinc binding could affect AP confor- mation as assayed by migration upon gel-filtration chromatog- raphy. Major AP species believed to correspond to monomeric, dimeric, and polymeric forms were observed (Fig. 2u). Total concentrations of Zn2+ as low as 0.4 PM decrease recoverable AP eluting from the column when compared with the elution pro- ftle obtained in the presence of EDTA and other metals (Fig. 2, a and b). At 25 p total ZnZ+, ~ 2 0 % of the A@ elutes. This deficit mainly affects the high order polymer and dimeric species which apparently do not enter the gel. Meanwhile, the relative amount of monomeric AB is preserved. A systematic assessment of several metals indicates that the reduction ofAP recoverable by chromatography is most sensitive to ZnZ+, with related tran- sition metals Co2+, Ni2+, and Fez+ (at 25 PM) displaying similar effects on chromatography to those obtained with only 10 PM ZnZ+ (Fig. 26). Other transition metals, heavy metals, and A13+ (25 PM) have partial effects on AP solubility comparable with 3 p~ total Zn'". Meanwhile, Ba2+, A$+, Mgz+, and Ca2+ (25 PM) have the least effect on AP compared with the EDTA profile, although 40% less total peptide appears to elute. Pb2+ (25 p ~ ) most strongly promotes the elution of the monomeric peptide, abolishing high order polymers; overall recovery is similar to that obtained with 0.4 PM total Zn2'. In making comparisons of the effects of these metal ions, it is again important to consider the differential metal ion chelating effects of Tris mentioned earlier.

A dramatic increase in AP dimerization is observed with Cu2+ (25 p~ total). This metal also induces exaggerated Ap absor- bance (4-fold) at 254 nm when compared with 214 nm absor- bance and induces the monomeric species to apparently fluo- resce at 254 nm causing negative readings (Fig. 2a) which are

12154 Precipitation of AP by Zinc

b

a Bound Zn(ll)/free 61

/ KA= 5.2 pM

0 0.5 1 1.5 2 2.5 Bound zn(ll) (pM)

'None 'AI(III) 'Fe(II) 'Ca(ll) 'Ag(II) 'Mg(ll) 'Fe(ll1) 'Cd(ll) 'Hg(ll) 'Cu(ll) in ( l I)*'Co(ll) 'Ni(ll) 'Pb(ll) 'Ba(II) 'Zn(ll) '

C Zn(ll) bound (96 Inmol) 100

70

d Bound Zn(ll)/iree

1.6

1.4-

1.21

60 - 1 - 50 - 40 - 30 -

0.611d

20 - 10 -

0

0.8 - f KA= 334 nM

- I

I I

achain Insulin Apmtinin BSA 1-17 12-28 25-35 1-28 1-28 40.1 1.40

Glnll 0 1 2 3 4 5

Bound Zn(ll) (WM)

e Yo Bmax

1M)

90

80

70

60

50

40

30

20

10

0 3 3.5 4 4.5 5 5.5 6 6.5 7 7.5

pH

FIG. 1. Analyses of esZnZ+ binding to Ap. Values shown are means -c S.D., n 2 3. a , Scatchard plot. Aliquots ofAP were incubated (60 min) with 65Zn2+ in the presence of varying concentrations of unlabeled Zn2+ (0.01-50 PM total). The proportion of 65Zn2+ binding to immobilized peptide (1.0 nmol) described two binding curves as shown. The high-affinity binding curve has been corrected by subtracting the low-affinity component, and the low-affinity curve has had the high-affinity component subtracted. b, specificity of the Zn2+ binding site for metals. AP was incubated (60 min) with "Zn2' (157 nM, 138,000 cpm) and competing unlabeled metal ions (50 PM total). c, "Zn2' (74 nM, 104,000 cpm) binding to negative (aprotinin, insulin a-chain, reverse peptide 40-1) and positive (bovine serum albumin (BSA)) control proteins and AD fragments (identified by their residue numbers within the AD sequence,glnll refers toAp,_,, where residue 11 is glutamine). Percent binding of total counts 65Zn2+/min added is corrected for the amounts (in nanomoles) of peptides adhering to the membrane. d, as for a, with AP,,, peptide substituting for AP,,,. 157 rn 65Zn (138,000 cpm) is used in this experiment to probe immobilized peptide (1.6 nmol). e , pH dependence of 65Zn2+ binding to AP,,,.

proportionally positive a t 214 nm (Fig. 2b). A higher concentra- tion of Cu2+ (80 J ~ M total) promotes increased recovery of AP, indicating that the presence of Cu2+ favors solubility in this system.

The metal ions which most favored AP solubility (Mg", 25 VM

and total Cu2+, 25 PM) were tested for their ability to stabilize AP in a soluble state in the presence of 25 1.1~ total Zn". These combinations neither rescue nor worsen Zn2+-induced loss ofAP recovery (Fig. 26) . Overall, these data suggest that Zn2+ binding

reduces the recovery of AP, whereas a chelating agent attenu- ates this effect.

To determine whether the zinc-induced loss of AP during chromatography occurs because of AP precipitation onto a sur- face of the apparatus, we attempted to block the effect. Pre- treating the column with 3% BSA as an adsorption blocker significantly increases the amounts of AP recovered from the column, indicating that the peptide precipitates onto a column component (Fig. 2c) . Blocking the column results in a 200%

Precipitation of AP by Zinc 12155

... ... .&(ti) 0.4 pM

-.,-.. Zn(l1) 25pM

-'*.-** Cu(l1) 25 pM

a 214 nm ab . Sepharoae SN

C

0 5 10 15 20 25 30 35 40 45 p.ptld. Lug>

FIG. 2. Effect of Zn2+ and other metals on Afi polymerization using G50 gel filtration chromatography. Resufts shown are indica- tive of n e 3 experiments where 55 pg ofAp is applied to the column and

AB in the presence of EDTA, 50 p ~ , ZnZ+, 0.4 y; Zn2*, 25 p ~ ; and Cu2+, eluted in 15 ml, monitored by 254 nm absorbance. a, chromatogram of

25 p ~ . The elution points of molecular mass standards and relative assignments of Ap peak elutions are indicated. Mass standards were blue dextran (2 x 106 kDa, V, = void volume), BSA (66 kDa), carbonic anhydrase (29 kDa), cytochrome c (12.4 kDa), and aprotinin (6.5 m a ) . The mass of AP is 4.3 kDa. b, relative amounts (estimated from areas under the curve) of soluble Ap eluted as monomer, dimer, or polymer in

Zn2+ or Cuz+ (the likelihood of Tris chelation is indicated by upper limit the presence of various metal ions 125 VM), varying concentrations of

estimates), and EDTA. Data for experiments performed in the presence of copper were taken from 214 nm readings and corrected for compari- son. e, effects of pre-blocking the chromatography column with BSA upon the recovery of AP species in the presence of zinc (25 y), copper (25 p"f, or chelator.

increase in the recovery of A@ in the presence of Zn2+ (25 VM total), a 75% increase in recovery in the presence of Cu2+ (25 VM total), but only a 10% increase in the presence of EDTA (50 VM). This confirms that precipitation onto the column is most spe- cifically accelerated by zinc.

To determine the part of the column onto which AP was precipitating, we incubated A@ solutions with various column components and assayed A@ concentrations by UV absorption

zn 2+ cu *+ EDTA

b Remainder in supernatant (%)

1

20

15

10

5

0 zn *+ cu EDTA

FIG. 3. Ap binding to kaolin (aluminum silicate): effects of zinc (25 PM), copper (25 p ~ ) , and EDTA (50 p ~ ) . a , concentration (by 214 nm absorbance) of Aj3 remaining in supernatant after incubation with 10 mg of G50 Sephadex. 6 , concentration (by 214 nm absorbance,) ofAp remaining in supernatant after incubation with 10 rng of kaolin, ex- pressed as percent of the starting absorbance.

before and after the incubation. Replicating the chromatogra- phy experimental conditions, A@ (100 PM in equilibration buffer) was incubated for 1 h in plastic reaction vessels with or without the presence of Sephadex. Loss to the plastic accounts for 45% of the observed precipitation, to siliconized plastic <l%, and binding to Sephadex <1% (data not shown). Hence, A@ precipitates are unlikely to be adsorbing to the Sephadex or plastic support. However, similar incubations in borosilicate glass test tubes result in 20% adsorption, which increase to 35% in the presence of zinc (25 p~).

The glass in the Bio-Rad Econo Columns is made of 7740 Pyrex (Corning, Park Ridge, IL) and is composed of SiO,, 80.6%; B,O,, 13.0%; Na,O, 4.0%; and A1,0,, 2.3%. Because of reports associating alumi~osilicates with @-amyloid deposition (Masters et al., 1985a; Candy et a€., 1986), we proceeded to test whether A@ binds to aluminum silicate. We observed rapid and extensive binding of A@ to kaolin, an insoluble hydrated alu- minum silicate. Moreover, incubation of A@ (0.4 mgfml) with Sephadex (5%, v/v) in the presence of zinc, copper, or EDTA causes only small changes in solubility which may be attrib- uted to binding to the plastic in the reaction vessels (Fig. 3a). Incubation of A@ (0.4 mg/ml) with kaolin (5%, v/v, 5 min, room temperature), causes precipitation of up to 87% of the peptide present. This precipitation is greatest in the presence of zinc

12156 Precipitation of AP by Zinc a

0.1 10 25 50 733 Conc. Zn 2t (lotla. pM)

FIG. 4. Effect of Zn2’ upon AP resistance to tryptic digestion. a, a blot of tryptic digests of AD (13.9 pg) after incubation with increasing concentrations of zinc (lane labels, in micromolar), stained by Coo- massie Blue. Digestion products of 3.6 kDa (AD,,), and 2.1 kDa (A&,,,), as well as undigested AP,,, (4.3 kDa), are indicated on the left. The migration of the low molecular size markers ( S T D ) are indicated (in kilodaltons) on the right. h, “Zn2* binding to AB tryptic digestion products. The blot in a was incubated with “‘Zn“’, the visible bands excised, and the bound counts for each band determined. These data are typical of n = 3 replicated experiments.

(25 PM) where the amount of AP recovered from the zinc incu- bation supernatant is nearly half of the amount recovered from the EDTAincubation supernatant (Fig. 36). The effect of copper (25 PM) upon kaolin-induced AP precipitation is similar to the effect of EDTA (Fig. 3b). The binding of AP to kaolin is not reversible to subsequent treatment with 10 mM EDTA, but can be eluted by 2 M NaOH (data not shown).

To further test whether zinc induces irreversible precipita- tion ofAP in the absence of kaolin, AP incubated with Zn2+ (200 PM, 1-24 h, 20 “C) was subjected to SDS Tris/Tricine gel elec- trophoresis. The monomeric species was the major band de- tected on Coomassie-stained gels and migrated identically to unincubated AP (data not shown), indicating that zinc does not induce covalent or SDS-resistant polymerization of AP.

Since the APP secretase site at Lys-16 (Esch et al., 1990; Sisodia et al., 1990) in AP is within the obligatory zinc binding region, we next tested the ability of Zn2+ to protect AB from secretase-type cleavage by trypsin, a serine-protease whose ac- tivity was found to be unaffected by zinc. Amino-terminal se- quence on AP tryptic digestion products transferred to polyvi- nylidene difluoride membrane following SDS-polyacrylamide gel electrophoresis indicated two detectable fragments corre- sponding to residues 6-40 and 17-40 (Fig. 4a). The predicted tryptic cleavage product representing residues 29-40 did not appear on the blot and may not be retained by the polyvinyli- dene difluoride membrane during transfer and treatment. Di- gestion is inhibited by the presence of increasing concentra- tions of Zn2’. At 200 PM, Zn2’ causes complete inhibition of AP hydrolysis; however, at this zinc level, tryptic activity is also slightly inhibited. Probing the blot with “Zn2+ confirmed the zinc binding identity of the peptide fragments and facilitated quantification of the hydrolysis of the zinc binding site (Fig. 4b). The rate of digestion of AS,-,, and the AB,, fragment is inhibited by the presence of zinc, whereas the digestion of the

fragment is not inhibited by increasing zinc concentra- tions. Hence, only the peptides possessing the intact zinc bind- ing domain of AP (residues 6-28), and therefore capable of binding Zn2’ (Fig. 4b), have their rates of digestion inhibited by zinc in this experiment. These data indicate that secretase-type

cleavage of AP can be inhibited by Zn2’ binding to the AP sub- strate.

DISCUSSION

Our data indicate that soluble AP,-,, possesses high and low affinity zinc binding affinities. The zinc binding site on AP maps to residues 6-28, with possibly conformational- and his- tidine-dependent properties. The affinity constants for zinc binding indicate that both binding associations are within physiological zinc concentrations, but that occupancy of the low affinity binding site may be associated with accelerated pre- cipitation ofAP by aluminum silicate (kaolin). occupancy of the high affinity site appears to have little effect on AP precipita- tion and is very highly specific, although our data cannot ex- clude the possibility of specific binding sites for alternative metals elsewhere on AP. Copper’s strong conformational inter- action (dimerization and fluorescence) with AP indicates that it may also directly interact with the peptide and may have a role in preventing AP precipitation onto aluminum silicate.

Extracellular zinc may play a role in the physiology of APP function by modifying its adhesiveness to extracellular matrix elements (Bush e f al., 1993). This is important because APP may play a role in cell adhesiveness (Shivers et al., 1988) and neurite outgrowth (Milward et al., 1992). The physiological function of the AS-zinc interaction remains unclear; however, increased resistance of AP to proteolytic cleavage in the pres- ence of zinc would increase the peptide’s biological half-life, and increased adhesiveness may also promote its binding to extra- cellular matrix elements. It has been reported recently that AP promotes neurite outgrowth by complexing with laminin and fibronectin in the extracellular matrix (Koo et al., 1993). Hence, both APP and AP may interact with the extracellular matrix to modulate cell adhesion. The possibility that zinc is a local en- vironmental cofactor modulating this interaction merits fur- ther investigation.

APP is highly abundant in platelets and brain (Bush et al., 1990) where zinc is also highly concentrated (Baker et al., 1978; Frederickson, 1989). Although APP is concentrated in vesicles in both of these tissues (Bush et al., 1990; Schubert et al., 1991), and zinc is actively taken up (Wolf et al., 1984) and stored in synaptic vesicles in nerve terminals throughout the telen- cephalon (Perez-Clause11 and Danscher, 1985), the colocaliza- tion of APP with zinc in these vesicles has yet to be demon- strated. Vesicular zinc storage is thought to play a role in stabilizing functional molecules such as NGF and insulin as insoluble intravesicular precipitates (Frederickson et al., 1987). Zinc may similarly play a role in stabilizing APP and AP.

The interaction between AP and zinc may be compared with that of insulin, a peptide whose zinc binding properties are well characterized. Like AP, insulin exhibits histidine-dependent high-affinity (KO = 5 PM) and low-affinity (KO = 140 p ~ ) zinc binding with stoichiometries of 1:l (insu1in:zinc) and 1:2, re- spectively (Goldman and Carpenter, 1974). Additionally, metal- free insulin exhibits a pH-dependent polymerization pattern consisting of monomer, dimer, tetramer, hexamer, and higher aggregation states, in dynamic equilibrium. At neutral pH, zinc and other divalent metal ions shift the equilibrium toward the higher aggregation states. At stoichiometric ratios of Zn’+:insulin in excess of 0.33, the peptide precipitates (Fred- ericq, 1956), reminiscent of zinc’s effects upon AP observed in the current studies.

AP chelates zinc with such high affinity that reports of its neurotoxic effects in neuronal cultures (Yankner et al., 1990; Koh et al., 1990) might be explained by a disturbance of zinc homeostasis. AP accumulates most consistently in the hip- pocampus, where extreme fluctuations of zinc concentrations occur (0.15-300 VM) (Frederickson, 1989), e.g. during synaptic

P ~ e c i p ~ ~ ~ ~ i o n

transmission (Assaf and Chung, 1984; Howell et al., 1984; Xie and Smart, 1991). Choi and co-workers (Weiss et al., 1989) have proposed that this trans-synaptic movement of zinc may have a normal signaling function and may be involved in long term potentiation. The hippocampus is the region of the brain that both contains the highest zinc concentrations (Frederickson et al., 1983) and is most severely and consistently affected by the pathological lesions of Alzheimer’s disease (Hyman et al., 1986). One of the prominent neurochemical deficits in Alzhei- mer’s disease is cholinergic deafferentation of the hippocam- pus, which has been shown to raise the concentration of zinc in this region (Stewart et al., 1984).

The rapid zinc-accelerated precipitation of AB by aluminum silicate (kaolin) is significant because of the candidacy of alu- minum as a pathogenic agent in AD (Per1 and Brody, 1980). Recent reports of Zn2’- and AI3+-induced sedimentation of A@ (Mantyh et ai., 3.9931, and the nucleation ofAp precipitation by aluminosilicate (Candy et al., 1992) also support our current observations.

Evidence for altered zinc metabolism in AD includes de- creased temporal lobe zinc levels ( W e n s t ~ p et al., 1990; Con- stantinidis, 1990; Corrigan et al., 19931, elevated (80%) cere- brospinal fluid levels (Hershey et al., 1983), increased hepatic zinc with reduced zinc bound to metallothionein (Lui et al., 19901, a Zn~+-modulated abnormality of APP in AD plasma (Bush et al., 1992), an increase in extracellular Zn*+-metallo- proteinase activities in AD hippocampus (Backstrom et al., 1992), and decreased levels of astrocytic growth inhibitory fac- tor, a metallothionein-like protein which chelates zinc (Uchida et aE., 1991). Collectively, these reports indicate that there may be an abnormality in the uptake or distribution of zinc in the AD brain causing high extracellular concentrations and low intracellular concentrations in the brain. Meanwhile, environ- mentally induced elevations of brain concentrations of both zinc (Duncan et al., 1992) and aluminum (Garruto et al., 1984; Per1 et al., 1982) have been implicated in the pathogenesis of Guamanian amyotrophic lateral sclerosislParkinson’s demen- tia complex, a disease also characterized by neurofibrillary tangles (Guiroy et a[., 1987). Interestingly, a pervasive abnor- mality of zinc metabolism manifested by immunological and endocrine dysfunction has been described as a common compli- cation of Down syndrome (Franceschi et al., 1988; Bjorksten et al., 3.9801, a condition characterize^ by the invariable onset of presenile A@ deposition and Alzheimer’s disease (Rumble et al., 1989).

Our results indicate that abnormally high zinc concentra- tions increase AP resistance to secretase-type cleavage and also accelerate Ap precipitation onto aluminosilicates. Zinc-induced accumulation of Ap in the neuropil may, in turn, invoke a glial inflammatory response, free radical attack, and oxidative cross-linking to form an, ultimately, “mature” amyloid. Collec- tively, these findings support the biochemical rationale for the chelation approach in the therapy ofAlzheimer’s disease (Crap- per McLachlan et al., 1991), since reduction of cerebral concen- trations of both aluminum and zinc could potentially decelerate the precipitation of A@.

Acknowledgments-We thank Dr. Peter Lansbury, Department of Chemistry, MIT for critically appraising the manuscript and for con-

Multhaup, Colin Masters, and Konrad Beyreuther for helpful discus- tributing peptides and facilities, Drs. Wilma Wasco, John Maggio, Gerd

sions, and Dr. Richard Cook, Biopolymers Laboratory, MIT, for peptide sequencing.

REFERENCES

A s 4 S. Y., and Chung, S.-H. 119841 Nature 308,734-736 Backstrom. 3. R., Miller, C. A., and Tokes, 2. A. ( 19921 J. N e ~ r ~ h e m , 58,983-992 Baker, R. J., McNeil, J. J., and Lander, H. (1978) Thromb. Huemostasis 39, 360-

365

12157 Bjorksten, B., Back, O., Gustavson, K. H., Hallmans, G., Hagglof, B., and Tarnvik,

A. (1980) Acta Paediatr. Scand. 69, 183-187 Bush, A. I., Martins, R. N., Rumble, B., Moir, R., Fuller, S., Milward, E., Cume, J.,

Ames, D., Weidemann, A., Fischer, P., Mul th~up, G., Beyreuther, K., and Mas-

Bush,A. I., Whyte, s., Thomas, L. D.. Williamson, T. G., Van Tiggelen, C. J., Currie, ters, C. L. (1990) J . Biol. Chem. 265, 15977-15983

J., Small, D. H., Moir, R. D., Li, 9.-X., Rumble, B., Monning, U., Beyreuther, K., and Masters, C. L. (1992) Ann. Neurol. 32, 57-65

Bush, A. I., Multhaup, G., Moir, R. D., Williamson, T. G., Small, D. H., Rumble, B., Pollwein, P., Beyreuther, K., and Masters, C. L. 11993) J. Biol. Chem. 268,

Candy, J. M., Klinowski, J . , Perry, R. H., Perry, E. K., Fairbairn, A,, Oakley, A. E., 16109-16112

Carpenter, T. A,, Atack, J. R., Blessed, G., and Edwardson, J . A. (19861 Lancet 1,354-357

Candy, J. M., Oakley, A. E., and Edwardson, J. A. (1992) Biochem. Soc. ’Pans. 21, 53S.iAbstractI

Constantinidis. J. (1990) Encephak IS, 231-239 Corrigan, F. M., Reynolds, G. P., and Ward, N. I. (1993) Biometals 6, 149-154 Crapper McLachlan, D. R., Dalton, A. J. , Kruck, T. P. A,, Bell, M. Y., Smith, W. L.,

Kalow, W., and Andrews, D. F. (1991) Lancet 337, 1304-1308 Dawson, R. M. C., Elliott, D. C., Elliott, W. H., and Jones, K. M. (19861 Data for

~ i ~ h e m i c a ~ Research, Oxford University Press, Oxford Duncan, M. W.. Marini, A. M., Watters, R., Kopin, I. J., and Markey, S. P. (19921 J.

Neurosci. 12, 1523-1537 Esch, F. S., Keim, P. S., Beattie, E. C., Blacher, R. W., Culwell, A. R., Oltersdorf, T.,

McClure, D., and Ward, €? J. (1990) Science 248, 1122-1124 Franceschi. C., Chirieolo, M., Licastro, F., Zannotti, M., Masi, M.. Mocchegiani, E.,

and Fabris, N. (1988) J. Ment. Defic. Res. 32, 169-181 Frederickson, C. J. i 19891 Int. Reo. Neurnbiol. 31,145328 Frederickson, C. J., Witenick, M. A., Manton, W. I., and Kirkpatrick, J. B. (1983)

Frederickson, C. J., Perez-Clausell, J., and Danscher, G. (19871 J . Histochem.

Fredericq, E. 119561 Arch. Biochem. Biophys. 65,218-228 Garruto, R. M., Fukatsu, R., Yanagihara, R. T., Gadjusek, D. C., Hook, G., and

Fiori, C. E. (1984) Proc. Natl. Acad. Sci. U. S. A. 81, 1875-1879 Glenner, G. G., and Wong, C. W. (1984) Biochem. Biophys. Res. Commun. 120,

885-890 Goate, A,, Chartier-Harlin, M., Mullan. M., Brown, J., Crawford, F., Fidani, L.,

Giuffra, L., Haynes, A,, Irving, N., James, L., Mant, R., Newton, P., Rooke, K., Roques, P., Talbot, C., Pericak-Vance, M., Roses, A., Williamson, R., Rossor, M., Owen, M., and Hardy, J . (1991) Nature 349,704-706

Goldgaber, D., Lerman, M. I., McBride, 0. W., Saffiotti, U., and Gajdusek, D. C. (1987) Science 235,877-880

Goldman, J., and Carpenter, F. H. (1974) Biochemistry 13,45664574 Guiroy, D. C., Miyazaki, M., Multhaup, G., Fischer, P., Garruto, R. M., Be-yreuther, K., Masters, C. L., Simms, G., and Gibbs. C. J., Jr (1987 I Proc. Nazi Acad. Sei. U. S. A. 84,2073-2077

Haass, C., Schlossmacher, M. G., Hung, A. Y., Vigo-Pelfrey, C., Mellon, A,, Ostaszewski, B. L., Lieberburg, I., Koo, E. H., Schenk, D., Teplow, D. B., and

Hendriks, L., van Duijn, C. M., Cras, P., Cruts, M., van Nul, W., van Harskamp, F., Selkoe, D. J. (1992) Nature 359,322-325

Warren, A., McInnis, M. G., Antonarakis, S. E., Martin, J.-J., Hofman, A,. and van Broeckhoven, C. t 1992) Nature Genet. 1, 218-221

Hershey, C. O., Hershey, L. A., Varnes,A., Vibhakar, S. D., Lavin, P., and Strain, W.

Howell, G. A., Welch, M. G., and Frederickson, C. J. (1984) Nature 308, 736-738 H. (19831 Neurology 93, 1350-1353

Hyman, B. T., Van Hoesen, G. W., Kroner, L. J., and Damasio, A. R. (19861 Ann.

Kang, J., Lemaire, H., Unterheck,A., Salbaum, J. M., Masters, C. L., Grzeschik, K., Neurol. 20, 472481

Multhaup. G., Beyreuther, K., and Muller-Hill, B. (1987) Nature 326, 733-736 Koh, J., Yang, L. L., and Cotman, C. W. (1990) Brain Res. 633, 315-320 Koo, E. H., Park, L., and Selkoe, D. J. (1993) Proc. Natl. Acad. Sci. U. S. A. 90,

47484752 Lw, E., Carman, M. D., Femandez-Madrid, I. J., Power, M. D., Lieberburg, I., Van

Duinen, S. G.. Bots, G. T. A. M., Luyendijk, W., and Frangione, B. (1990) Science

Lui, E., Fisman, M., Wong, C., and Diaz, F. (1990) J . Am. Geriatr Soc. 38,633-639 248, 1124-1126

Mantyh, P. W., Ghilardi, J. R., Rogers, S., DeMaster, E.,Allen. C. J., Stimson. E. R.,

Masters, c. L., ~ u l t h a u p , G., Simms, G., Pottgiesser, J., Martins, R. N., and

Masters, C. L., Simms, G., Weinman, N. A., Multhaup, G., McDonald, R . L., and

Milward, E. A., Papadopoulos, R., Fuller, S. J., Moir, R. D., Small, D., Beyreuther,

Murrell, J., Farlow, MI.. Ghetti, B., and Benson, M. D. (1991) Science 254,97-99 Perez-Clausell, J.. and Danscher, G. 11985) Bruin Res. 337,91-98 Ped, D., and Brody, A. (1980) Science 208, 297-299 PerL D. P., Gajdusek, D. C . . Garruto, R. M., Yanagihara, R. T., and Gibbs, C. J.

Robakis, N. K., Ramakrishna, N., Wolfe, G., and Wisniewski, H. M. (1987) Proc. (19821 Science 217, 1063-1055

Rumble, B.. Retallack, R., Hilbich, C.. Simms, G., Multhaup, G., Martins, R., Nail. Arad. Sci. U. S. A. 84, 41904194

Hockex A., Montgomery. P., Beyreuther, K., and Masters, C. L. (1989) N. Engl. J . Med. 320, 1446-1452

Schubert, W., Prior, R., Weidemann,A., Dircksen, H., Multhaup, G., Masters, C. L., and Beyreuther, K. (1991) Brain Res. 563, 184-194

Seubert, p.. Kgo-Pdfrey. C., Esch, F., Lee, M., Dovey, H., Davis, D., Sinha, S., Schlossmacher, M., Whaley, J., Swindlehurst, C., MeCormack, R., Wolfert, R.,

Shivers, B. D., Hilbich, C . , Multhaup, G., Salbaum, M., Beyreuther, K., and See- SeIkoe, D., Lieberburg. I. , and Schenk, D. (1992) Nature 369, 325-327

Brain Res. 273, 335-339

Cytochem. 35,579-583

and Maggio, J. E. (1993) J. Neurochem. 61, 1171-1174

Beyreuther, K. (1985al EMBO J. 4,2757-2763

Beyreuther, K. (1985b) Proc. Natl. Acad. Sci. U. S. A. 82, 42454249

K., and Masters, C. L. (1992) Neuron 9, 129-137

12158 P r e e i ~ i ~ ~ t i o n of AP by Zinc

Shoji, M., Golde, T. E., Ghiso, J.. Cheung, T. T., Estus, S., Shaffer, L. M., Cai, X.-D., burg, P. H. (1988) EMBO J . 7, 1365-1370

McKay, D. M., Tintner, R., Frangione, B., and Younkin, S. G. ( 19921 Science 258, 126-129

Sisodia, S. S., Koo, E. H., Beyreuther, K., Unterbeck, A,, and Price, D. L. (1990) Science 248,492-495

Stewart, G. R., Frederickson, C. J., Howell. G. A,, and Gage, F. H. (1984)Brain Res. 290,4331

Tanzi, R. E., Gusella, J. E, Watkins, P. C., Bruns, G.A. P., St George-Hyslop, P., van Keuren, M. L., Patterson, D., Pagan, S., Kurnit, D. M., and Neve, R. L. (1987)

Tomski, S., and Murphy. R. M. (1992) Arch. Biocehem. Biophys. 294,630-638 Uchida, Y., Takio, K., Titani, K., Ihara, Y., and Tomonaga, M. (1991) Neuron 7,

Weiss, J. H., Koh, J., Christine, C . W., and Choi, D. W. (1989) Nature 338, 212 Wenstrup, D., Ehmann, W. D., and Markesbery, W. R. (1990) Brain Res. 533,

Wolf, G., Scutte, M., and Rtimhild, W. (19841 Neurosci. Lett. 51, 277-280 Xie, X., and Smart, T. G . (19911 Nature 349,521-524 Yankner, ti. A., Duffy, L. K., and Kirschner, D. A. (1990) Science 260,279-282

Science 235, 880-884

337-347

125--131