Cathepsins S, B and L with aminopeptidases display β-secretase activity associated with the pathogenesis of Alzheimer's disease

Article abstract of DOI:10.1515/BC.2011.054

β-site APP-cleaving enzyme (BACE1) cleaves the wild type (WT) β-site very slowly (k_cat/K_m: 46.6 m^-1 s^-1). Therefore we searched for additional β-secretases and identified three cathepsins that split the WT β-site much faster. Human cathepsin S cleaves the WT β-site (k_cat/K_m: 54 700 m^-1 s^-1) 1170-fold faster than BACE1 and cathepsins B and L are 440- and 74-fold faster than BACE1, respectively. These cathepsins split two bonds flanking the WT β-site (K-MD-A), where the K-M bond (85%) is cleaved more efficiently than the D-A bond (15%). Cleavage at the major K-M bond yields Aβ (amyloid β-peptide) extended by N-terminal Met that should be removed to generate Aβ initiated by Asp1. The activity of cytosol and microsomal aminopeptidases on relevant peptides revealed rapid removal of N-terminal Met but not N-terminal Asp. Brain aminopeptidases showed similar specificity. Thus, aminopeptidases would convert Aβ extended by Met into regular Aβ (Asp1) found in amyloid plaques. Earlier studies indicate that Aβ is likely produced in the endosome and lysosome system where cathepsins S, B and L are localized and cysteine cathepsin inhibitors reduce the level of Aβ in cells and animals. Taken together, cathepsins S, B and L deserve further evaluation as therapeutic targets to develop disease modifying drugs to treat Alzheimer's disease.

Full text of DOI:10.1515/BC.2011.054

Article in press - uncorrected proof
Biol. Chem., Vol. 392, pp. 555–569, June 2011 • Copyright ꢀ by Walter de Gruyter • Berlin • New York. DOI 10.1515/BC.2011.054
Cathepsins S, B and L with aminopeptidases display
b-secretase activity associated with the pathogenesis
of Alzheimer’s disease
would slow down, stop or reverse the progression of AD.
The most successful clinical application of this approach are
inhibitors of the HIV protease to treat AIDS and inhibitors
of the angiotensin-converting enzyme to treat hypertension
and congestive heart failure. Approximately 10 years ago,
the b-site APP-cleaving enzyme (BACE1) was identified as
b-secretase (Hussain et al., 1999; Sinha et al., 1999; Vassar
et al., 1999; Yan et al., 1999; Lin et al., 2000) and the
enzyme has been intensively studied (Wolfe, 2006; Cole and
Vassar, 2008). Potent BACE1 inhibitors are available (Abbe-
nante et al., 2000; Turner et al., 2001) but progress in drug
development to treat AD is rather slow (Cole and Vassar,
2008). Therefore, alternative approaches have been evaluated
(Hook et al., 2008b). It was shown that cathepsin B can
function as b-secretase and that cysteine protease inhibitors
improve memory and reduce amyloid plaques in the brain
(Hook et al., 2005, 2008a). The glutaminyl cylase mediates
the formation of pyroglutamated truncated Ab (pGlu3,
Ab3–42) that facilitates Ab aggregation, and inhibitors of glu-
taminyl cyclase reduce plaque burden (Schilling et al., 2006,
2008).
Compelling evidence suggest that BACE1 is the genuine
b-secretase. Five research groups identified BACE1 as the
b-secretase by using different strategies: cloning expression
(Vassar et al., 1999), expressed sequence tag (EST) (Hussain
et al., 1999; Lin et al., 2000), analysis of Caenorhabditis
elegans genome and human EST (Yan et al., 1999), and
affinity chromatography (Sinha et al., 1999). All research
teams assumed that the b-secretase should be an aspartic
protease with optimal activity at acidic pH, cleaving the b-
site of the Swedish (SW) mutant sequence faster than the b-
site of the wild type (WT) sequence. Accordingly, in all cases
b-secretase activity was monitored by peptide substrates con-
taining the b-site SW sequence and for affinity purification,
the immobilized ligand was also derived from the b-site SW
sequence. The presumed properties of the b-secretase were
mainly based on the activity of brain extracts on SW and
WT peptides. We have shown that the specificity of cathep-
sin D toward SW and WT peptides is similar to that of
BACE1 and that the apparent b-secretase activity in brain
extracts is mainly due to cathepsin D, not BACE1 (Schechter
and Ziv, 2008). Additional problems with BACE1 are out-
lined below.
Israel Schechter* and Etty Ziv
Yenuvun Research Laboratory, 19 Hayetsira St., Rehovot
76380, Israel
*Corresponding author
e-mail: israels@yenuvun.com
Abstract
b-site APP-cleaving enzyme (BACE1) cleaves the wild type
(WT) b-site very slowly (kcat/Km: 46.6 M
-1 s-1). Therefore we
searched for additional b-secretases and identified three
cathepsins that split the WT b-site much faster. Human
cathepsin S cleaves the WT b-site (kcat/Km: 54 700
-1
M
s-1)
1170-fold faster than BACE1 and cathepsins B and L are
440- and 74-fold faster than BACE1, respectively. These
cathepsins split two bonds flanking the WT b-site (K-MD-
A), where the K-M bond (85%) is cleaved more efficiently
than the D-A bond (15%). Cleavage at the major K-M bond
yields Ab (amyloid b-peptide) extended by N-terminal Met
that should be removed to generate Ab initiated by Asp1.
The activity of cytosol and microsomal aminopeptidases on
relevant peptides revealed rapid removal of N-terminal Met
but not N-terminal Asp. Brain aminopeptidases showed sim-
ilar specificity. Thus, aminopeptidases would convert Ab
extended by Met into regular Ab (Asp1) found in amyloid
plaques. Earlier studies indicate that Ab is likely produced
in the endosome and lysosome system where cathepsins S,
B and L are localized and cysteine cathepsin inhibitors
reduce the level of Ab in cells and animals. Taken together,
cathepsins S, B and L deserve further evaluation as thera-
peutic targets to develop disease modifying drugs to treat
Alzheimer’s disease.
Keywords: BACE1; cleavage of b-site sequence; cleavage
points; kcat; Km; new b-secretases; WT b-site sequence.
Introduction
Many studies indicate that the amyloid b peptide (Ab),
deposited in the brain as amyloid plaques, plays a major role
in the pathogenesis of Alzheimer’s disease (AD) (Selkoe,
2001). The Ab peptide is generated from the amyloid pre-
cursor protein (APP) by the consecutive action of two pro-
teases: b-secretase and then g-secretase. Major efforts have
been directed to identify the secretases because inhibitors
that block their activity reduce the formation of Ab and thus
BACE1 cleaves the mutated SW b-site sequence of APP
present in two Swedish families much better than the WT
sequence present in the worldwide population. BACE1
cleaves the WT b-site very slowly (kcat/Km: approx. 50 M
-1 s-1,
Lin et al., 2000; Shi et al., 2001; Schechter and Ziv, 2008)
2010/313
Brought to you by | University of California - San Francisco
Authenticated
Download Date | 12/13/14 1:15 AM

Article in press - uncorrected proof
556 I. Schechter and E. Ziv
whereas proteases acting on relevant substrates are much
more efficient, kcat/Km: 10 000–1 000 000
-1 s-1 (Schechter,
progressed (see below). Therefore, we analyzed bovine kid-
ney because the APP is expressed ubiquitously (Neve et al.,
1988) and many cell types, including kidney cells, produce
Ab (Vassar et al., 1999).
M
1970; Dunn and Hung, 2000). BACE1 is a membrane bound
enzyme and it was argued that enzymes anchored in the cell
membrane could have reduced activity when freed in solu-
tion. However, application of the active site mapping pro-
cedure (Schechter and Berger, 1967) to evaluate the
specificity of BACE1 have led to the synthesis of good
peptide substrates that differ from the WT b-site sequence
(Turner et al., 2001; Tomasselli et al., 2003; Shi et al., 2005).
For example, a peptide that extensively differs from the WT
b-site octapeptide sequence (P4-P49) in seven out of eight
The 25V fraction of bovine kidney was chromatographed
on a G-100 column and b-secretase activity was monitored
by mixing aliquots of the eluate with the WT peptide sub-
strates P4 or P17 (see Figure 1 for sequences of the peptide
substrates and cleavage products). Peptide degradation was
observed in two fractions (Figure 2). A fraction of high
molecular mass (G70 kDa, between the void volume and
BSA) yielded multiple degradation products of varying size
likely due to a mixture of proteases, and it was not further
investigated. The other fraction of low molecular mass
(approx. 22 kDa, near the myoglobin size marker) showed a
distinct peak of activity and a simple pattern of cleavage
products, indicating for activity by a single protease. The
cleavage products of P4 and P17, observed by paper electro-
phoresis, were consistent with cleavage at the b-site of these
peptide substrates. It was previously shown that BACE1 and
cathepsin D cleave the P4 WT peptide at the M-D b-site to
yield the P6 C-terminal peptide wsubsequently it was found
that the peptide product is P6qM and not P6 that have sim-
ilar electrophoretic mobility (see Figure 6)x, and the P9 N-
terminal peptide (migrates like P4 on paper electrophoresis,
Schechter and Ziv, 2008), as observed in Figure 2A. It was
also shown (Schechter and Ziv, 2008) that cleavage at the
M-D b-site of P17 yields the N wsubsequently it was found
that the peptide is N-M and not N that has similar electro-
phoretic mobility (see Figure 5)x and C terminal peptide
products with electrophoretic mobility like that observed
(Figure 2B). Therefore, the active fractions (Figure 2, frac-
tions 32–36) were pooled, concentrated and loaded on CM52
cation exchanger. The b-secretase activity was found in the
effluent. It was concentrated and loaded on a gel made in
neutral buffer. After electrophoresis the gel was sliced into
amino acid residues, is cleaved by BACE1 with kcat/Km of
-1
342 000
M
s-1 (Turner et al., 2001). These findings suggest
that BACE1 acts on yet unknown substrate(s) and it is
unlikely that the WT APP is the physiological substrate
(Schechter and Ziv, 2008). Knockout of BACE1 in mice
markedly reduces Ab formation (Cai et al., 2001; Lou et al.,
2001; Roberds et al., 2001). Yet, it was shown that knockout
of other genes in rodents, and the corresponding genetic
defects in human, might reveal different phenotypes. Nine
different tumor suppressor genes associated with human
tumors were mutated in mouse and rat but only two mutated
genes overlapped with the human phenotype (Jacks, 1996).
Considering these issues, we searched for other b-secretase
candidate(s). We found cathepsin D and evaluated properties
of cathepsin D related to BACE1 that were not examined
before. It was found that human BACE1 and cathepsin D
display the same specificity; they have similar kinetic con-
stants (kcat, Km, kcat/Km) for cleaving b-sites of relevant pep-
tides, cleaving SW better than the WT sequence. G-100
chromatography of brain extract reveal that b-secretase
activity is eluted at the position of cathepsin D (42 kDa) and
not at that of BACE1 (72 kDa). Pepstatin A at 20 n
M com-
pletely inhibits the cleavage of SW peptides by both cathep-
sin D and brain extracts, whereas BACE1 activity on the
same peptides is not inhibited even by 1000-fold higher con-
centration of pepstatin A (20 mM). Quantitative Western
blots reveal that in human brain, cathepsin D is approxi-
mately 280-fold more abundant than BACE1. Therefore, it
is conceivable that b-secretase activity observed in brain
extracts is mainly due to cathepsin D and not BACE1
(Schechter and Ziv, 2008). Nevertheless, because both
BACE1 and cathepsin D show poor activity towards the WT
b-site sequence we continued the search for additional b-
secretase candidate(s). Here we report that cathepsins S, B,
and L cleave the WT b-site sequence much better than
BACE1 and cathepsin D.
Figure 1 Sequences of WT peptide substrates and of cleavage
products.
Substrates (P17, P4) match with the amino acid sequence flanking
the b-site of the WT APP, except for the carboxyl-terminal residue
of the P4 peptide where glycine is replacing histidine present in
APP. Cleavage products are designated with reference to cleavage
of the M-D bond of the WT b-site sequence. Accordingly, cleavage
of P17 (10-mer) can yield the N (N-terminal, 5-mer) and C (C-
terminal, 5-mer) peptides (cleavage at the b-site M-D bond), the N-
M (N minus Met residue) and CqM (C plus Met residue) peptides
(K-M bond split), the NqD (N plus Asp) and C-D (C minus Asp)
peptides (D-A bond split). Similarly, cleavage of P4 (14-mer) can
yield the P9 (8-mer) and P6 (6-mer) peptides (cleavage at the b-
site M-D bond), the P9-M (P9 minus Met) and P6qM (P6 plus
Met) peptides (K-M bond split), the P9qD (P9 plus Asp) and P6-
D (P6 minus Asp) peptides (D-A bond split).
Results
Search of b-secretase activity in the kidney reveals
cathepsin S
The 25V fraction of bovine brain homogenate was chroma-
tographed on a G-100 column and aliquots of the eluate were
reacted with the P17 WT peptide. A complex pattern of
cleavage products was observed that was resolved as work
Brought to you by | University of California - San Francisco
Authenticated
Download Date | 12/13/14 1:15 AM

Article in press - uncorrected proof
New b-secretase candidates rapidly cleave the WT b-site 557
Figure 2 b-Secretase activity of the 25V fraction of bovine kidney prepared in the pH 7.3 buffer and subjected to chromatography on
Sephadex G-100 at pH 7.3.
b-Secretase activity was tested by using P4 (A) and P17 (B) peptide substrates. Active fractions (32–36) were pooled for further purification.
Components of the reaction mixture were resolved by paper electrophoresis in the pH 6 buffer. Bovine serum albumin (BSA), ovalbumin
(OVA), myoglobin (MYO) and lysozyme (LYS) indicate the position of protein markers eluted from the G-100 column. Electrophoretic
mobility of the peptide substrates (bold P4 and P17) and of resolved major cleavage products are shown on the left.
strips and strong b-secretase activity was found in eluates of
strips one to three that were pooled (Figure 3A), concentrated
and loaded on a gel made in acidic buffer. The gel was sliced
and strong b-secretase activity was detected in eluates of
strips six and seven that were pooled (Figure 3B). Trichlo-
roacetic acid was added, the protein pellet was dissolved in
loading buffer and it was resolved by SDS-PAGE. The
stained gel showed three bands, two of which had molecular
masses of approximately 25 kDa and 29 kDa (Figure 3C).
Because G-100 chromatography reveals b-secretase activity
with a molecular mass of approximately 22 kDa, the 25 kDa
and 29 kDa bands were cut from the gel and digested by tryp-
sin. Tandem mass spectrometry revealed that the 25 kDa band
contained tryptic peptides derived from bovine cathepsin S.
or BACE1 (Schechter and Ziv, 2008). The latter enzymes
yield other HPLC peaks of peptide products formed by
cleavage at the M-D bond of the b-site sequence of the P17
or P4 substrates (Figure 1).
Mass spectroscopy analyses of HPLC peaks of P17 digest
by cathepsin S (Figure 4B) were as follows. The two major
peaks correspond to cleavage products at the K-M bond of
P17 to yield the N-M peptide (461 Da) and the CqM peptide
(767 Da). The two minor peaks correspond to cleavage prod-
ucts at the D-A bond of P17 to yield the NqD peptide (707
Da) and C-D peptide (521 Da). Thus, cathepsin S does not
split at the b-site M-D bond of the P17 WT substrate, but it
cleaves the WT sequence at two bonds flanking the WT b-
site: K-MD-A (hyphens indicate the bonds cleaved, Figure
1). From areas of the peaks (N-M plus CqM compared to
NqD plus C-D, six measurements) we estimate 85% cleav-
age at the K-M bond and 15% cleavage at the D-A bond of
P17. The resolved peaks (Figure 4B) were concentrated
approximately 25-fold and subjected to paper electrophoresis
(Figure 5). All cleavage products migrated similarly to the
appropriate peptide markers in complete agreement with
their mass as determined by mass spectrometry.
Cleavage of WT peptides by cathepsins S, L and B
Some peptide substrates and cleavage products are not
resolved by paper electrophoresis. Therefore, in order to
determine the site of substrate cleavage by cathepsin S, the
enzyme digest was further analyzed by HPLC and mass
spectroscopy. HPLC of digests of the P17 and P4 peptides
by cathepsin S revealed four new peaks (Figure 4B and H).
The HPLC patterns of P17 and P4 digests by cathepsin S
differ from those of the P17 and P4 digests by cathepsin D
Analyses of the P4 peptide digested by cathepsin S were
similar to those obtained with P17. HPLC (Figure 4H) and
mass spectrometry of resolved peaks showed: (1) two major
Brought to you by | University of California - San Francisco
Authenticated
Download Date | 12/13/14 1:15 AM

Article in press - uncorrected proof
558 I. Schechter and E. Ziv
Figure 3 Purification of b-secretase activity from bovine kidney by polyacrylamide gel electrophoresis.
(A) Neutral gel analysis. Effluent of the CM52 column with b-secretase activity was concentrated and subjected to electrophoresis on a
neutral polyacrylamide slab gel. After the run, the gel was sliced into twelve 3 mm wide strips that were soaked in elution solution. Aliquots
of the eluate were analyzed for proteins and b-secretase activity. SDS-PAGE developed with silver stain demonstrated proteins in the CM52
effluent (CM) loaded on the neutral gel and in eluates from strips 1–6 (counting from top of the neutral gel). The b-secretase activity
indicated on top of the gel was tested by using the P17 peptide substrate and paper electrophoresis. The extent of P17 degradation was
qualitatively graded between – (P17 not degraded) to qqqq (P17 completely degraded). Strong b-secretase activity was found in eluates
from strips 1–3 that were further analyzed. (B) Acidic gel analysis. Active fractions of the neutral gel (strips 1–3) were pooled, concentrated
(NE), resolved on an acidic gel and analyzed as above. Strong b-secretase activity was found in eluates of strips six and seven. (C) Active
fractions of the acidic gel (AC, stripes six and seven) were pooled, concentrated and resolved by SDS-PAGE developed with Coomassie
Blue. Asterisk indicates the 25 kDa band that contained tryptic peptides of bovine cathepsin S. Numbers indicate the molecular mass (kDa)
of protein size markers.
peaks corresponding to cleavage products of P4 at the K-M
bond to yield the P9-M peptide (832 Da) and the P6qM
peptide (824 Da) and (2) two minor peaks corresponding to
cleavage of P4 at the D-A bond to yield the P9qD peptide
(1078 Da) and P6-D peptide (578 Da). Here again, cathepsin
S cleaves the P4 WT peptide at two bonds flanking the WT
b-site: K-MD-A, as observed for the P17 peptide. From peak
areas (see above) we estimate that cleavage of P4 occurs
mainly at the K-M bond (approx. 90%) and less at the D-A
bond (approx. 10%), similar to the finding with P17. Cleav-
age products resolved by HPLC subjected to paper electro-
phoresis migrated similarly to the expected peptide markers
(Figure 6), in complete agreement with their mass deter-
mined by mass spectrometry.
The use of peptide markers, paper electrophoresis and
HPLC clearly identified the substrate and the cleavage prod-
ucts. Peptides (for example, P4 and P9-M) that were not
resolved by one procedure (paper electrophoresis, Figure 6)
Figure 4 HPLC profiles showing hydrolysis of the P17 and P4 peptide substrates by human cathepsins S, L and B.
(A) Control of P17 (1210 Da). (B) P17 with cathepsin S. Cleavage yields two major peaks of N-M (molecular mass: 461 Da) and CqM
(767 Da), and two minor peaks of NqD (707 Da) and C-D (521 Da). Molecular mass (Da) of the peptides was determined by mass
spectrometry of the resolved HPLC peaks. (C) Control of P17. (D) P17 with cathepsin L. HPLC profile is similar to that observed in panel
(B). Molecular mass values of peptides in resolved peaks are identical to those obtained in B. (E) Control of P17. (F) P17 with cathepsin
B. HPLC profile is similar to that observed in panels (B) and (D) except for markedly reduced CqM peak. Molecular mass values of
peptides in resolved peaks are identical to those obtained in panel (B). (G) Control of the P4 peptide substrate (1638 Da) and of three
markers of possible cleavage products. (H) P4 with cathepsin S. Cleavage yields two major peaks of P9-M (832 Da) and P6qM (824 Da),
and two minor peaks of P6-D (578 Da) and P9qD (1078 Da). Peptides were eluted from HPLC columns by 0.1% TFA in acetonitrile
gradients of 5–20% in (A, B, E, F); 5–25% in (C, D) and 10–25% in (G, H). Similar HPLC profiles and identical molecular mass values
of resolved peaks were obtained by analyses of the P17 and P4 WT peptide substrates digested by bovine cathepsins S, L and B.
Brought to you by | University of California - San Francisco
Authenticated
Download Date | 12/13/14 1:15 AM

Article in press - uncorrected proof
New b-secretase candidates rapidly cleave the WT b-site 559
Analyses of the P17 peptide digest by cathepsin B reveals
minor cleavage at the D-A bond, as seen from two small
HPLC peaks of NqD and C-D (Figure 4F) and confirmed
by mass spectrometry and paper electrophoresis. Cleavage at
the K-M bond is indicated by identifying the peptide pair,
N-M and CqM. However, the amount of the N-M peptide
is significantly higher than that of the CqM peptide as seen
by HPLC (Figure 4F) and paper electrophoresis (Figure 7A).
Conceivably this is due to the dual activity of cathepsin B,
as endopeptidase (cleaving the K-M and D-A bonds) and as
dipeptidyl carboxypeptidase (Aronson and Barrett, 1978).
The latter activity would degrade the CqM peptide by step-
wise removal from its C-terminus of the PheArg, AlaGlu and
MetAsp dipeptides (Figure 1). Indeed, the P17 digest by
cathepsin B reveals that the additional peptide spots co-
migrate with these dipeptide markers (Figure 7A). Thus, the
three cathepsins (S, L and B) exhibit similar specificity: a
major cleavage at the K-M bond (approx. 85%) and a minor
cleavage at the D-A bond (approx. 15%) flanking the WT
b-site (K-MD-A).
Cleavage products of P17 digest by cathepsin B is differ-
ent from that obtained by cathepsins S and L (Figure 7A)
due to the dipeptidyl carboxypeptidase activity of cathepsin
B. The cathepsin B cleavage pattern of P17 was observed in
the 25V fraction of bovine brain chromatographed on a G-
100 column that eluted at the molecular weight region of
Figure 5 Characterization of the P17 cleavage products produced
by cathepsin S.
P17 digested by human cathepsin S: total digest and fractions of
the total digest eluted from the HPLC column after 3 min, 13 min
and 23 min. Asterisks (13 min sample) indicate positions of the two
minor cleavage products (NqD top, C-D bottom) that are resolved
by HPLC. The positions of peptide markers are shown (amino acid
sequences of the markers are in Figure 1). Samples were resolved
by paper electrophoresis in the pH 6 buffer. Indicated on the left
are positions of the P17 peptide substrate (bold) and of the two
major cleavage products.
were clearly resolved by the other (HPLC, Figure 4H). Pep-
tide identification was further ascertained by mass spectrom-
etry. These analyses performed on peptide digests of
cathepsin S as well as of cathepsins L and B (see below)
enabled unambiguous identification of the bonds cleaved by
the enzymes.
In addition to cathepsin S, we studied cathepsins B and L
because high levels of these proteases were found in neurons
and amyloid plaques in the brain of Alzheimer’s patients by
enzyme cytochemistry and immunohistochemistry (Cataldo
and Nixon, 1990; Cataldo et al., 1991). Experiments con-
ducted with cathepsin L and the P17 peptide show cleavage
products similar to those observed with cathespsin S. Digests
of P17 by both enzymes have similar HPLC (compare Figure
4B and D) and paper electrophoresis (Figure 7A) patterns.
Mass spectrometry of HPLC peaks of P17 digest by cathep-
sin L had molecular mass values identical to those obtained
from HPLC of P17 digest by cathepsin S. Thus, cathepsin L
splits the WT P17 peptide at two bonds flanking the WT b-
site (K-MD-A), where the K-M bond (approx. 85%) is
cleaved more frequently than the D-A bond (approx. 15%),
as found for cathepsin S.
Figure 6 Characterization of the P4 cleavage products produced
by cathepsin S.
P4-digested by human cathepsin S: total digest and fractions of the
total digest eluted from the HPLC column after 7 min, 12 min and
14 min. Asterisks indicate positions of the two minor cleavage prod-
ucts (P6-D 7 min, P9qD 14 min). Samples were resolved by paper
electrophoresis in the pH 6 buffer. Indicated on the left are positions
of the P4 peptide substrate (bold) and of the two major cleavage
products.
Brought to you by | University of California - San Francisco
Authenticated
Download Date | 12/13/14 1:15 AM

Article in press - uncorrected proof
560 I. Schechter and E. Ziv
.
Figure 7 Characterization of the P17 cleavage products produced
by cathepsin B and by the 25V fraction of bovine brain subjected
to chromatography on a Sephadex G-100 column.
(A) Total digest of P17 by human cathepsin B, compared with the
P17 total digests by cathepsins S and L. Peptide markers are P17,
N-M and CqM peptides, the dipeptides MetAsp (MD), AlaGlu
(AE) and PheArg (FR). Samples were resolved by paper electro-
phoresis in the pH 6 buffer. Indicated on the left are positions of
the P17 peptide substrate (bold) and of the two major cleavage
products. Similar patterns of cleavage product were obtained by
paper electrophoresis analyses of P17 digested by bovine cathepsins
S, L and B. (B) 25V fractions of bovine brain were prepared as
described for bovine kidney. The 25V brain fraction prepared in the
pH 5.2 or pH 7.3 buffers were chromatographed on G100 equili-
brated at pH 5.2 or pH 7.3, respectively, and reacted with the P17
peptide. Reaction of G-100 fractions eluted in the range of 24–32
kDa (mass of cathepsin B is 28 kDa) are shown. Note that fractions
from the acidic G100 column (pH 5.2) cleave P17 to peptide prod-
ucts as observed for P17 cleavage by cathepsin B (see A), whereas
fractions from the neutral column (pH 7.3) do not degrade P17.
Indicated on the left are positions of the P17 peptide substrate
(bold), N-M and CqM peptides, the dipeptides MetAsp (MD),
AlaGlu (AE) and PheArg (FR).
parameters of these enzymes towards the P17 and P4 WT
peptide substrates (Table 1) by using paper electrophoresis.
The extent of hydrolysis of P17 was determined from the N-
M peptide spot because it originates from the major cleavage
site of P17 (approx. 85% cleavage of the K-M bond) and it
is well separated from other components of the reaction mix-
ture (Figures 2 and 5). Extent of hydrolysis of P4 by cathep-
sin S was determined from the P6qM peptide because it
originates from the major cleavage site of P4 (approx. 90%
cleavage at the K-M bond) and it is well separated from the
P4 substrate and other cleavage products (Figures 2 and 6).
Kinetic constants (kcat, Km, kcat/Km) were derived from Line-
weaver-Burk plots for cathepsins S and L.
Cathepsin B has endopeptidase activity that cleaves the
same bonds as cathepsins L and S. However, cathepsin B
has additional dipeptidyl carboxypeptidase activity that rap-
idly degrades the CqM cleavage product (90% in 1 h, data
not shown), it slowly degrades the N-M cleavage product
(5% in 1 h, data not shown) likely because N-M (tetrapep-
tide) is smaller than CqM (hexapeptide) (Schechter and Ber-
ger, 1967) and it probably degrades the C-terminus of the
P17 substrate. Therefore, in this case the P17 hydrolysis data
are problematic to construct Lineweaver-Burk plots to eval-
uate the kcat and Km constants. Instead, we estimated only
the kcat/Km value for P17 cleavage by cathepsin B. For this,
24–32 kDa (cathepsin B molecular mass is 28 kDa). Notably,
P17 degradation was observed in the brain 25V fraction pre-
pared in the pH 5.2 buffer and chromatographed on G-100
equilibrated with the pH 5.2 buffer but not in 25V fraction
prepared and chromatographed at pH 7.3 (Figure 7B). This
finding is consistent with cathepsin B activity in the brain,
because cathepsin B is stable at acidic pH but it is unstable
at neutral pH, whereas cathepsin S is stable at both acidic
and neutral pH (Bromme et al., 1993).
low P17 concentrations (0.4–1 mM) were used and the
released N-M peptide was determined. Apparent reaction
rates were first order with respect to the P17 concentration,
i.e., enzyme saturation was low. These values were used to
calculate kcat/Km, which represents the rate of hydrolysis at
infinite dilution of the substrate (Schechter, 1970).
The activity of bovine cathepsins S, L and B towards the
P17 and P4 peptides is similar to that observed for the cor-
responding human enzymes, as observed in the HPLC pro-
files (Figure 4), paper electrophoresis (Figure 7A) and kinetic
constants for cleaving the peptides (Table 1).
Kinetic constants for cleaving the WT peptides
by cathepsins S, L and B
To evaluate the relationship between the various cathepsins
and BACE1 in quantitative terms, we determined kinetic
Brought to you by | University of California - San Francisco
Authenticated
Download Date | 12/13/14 1:15 AM

Article in press - uncorrected proof
New b-secretase candidates rapidly cleave the WT b-site 561
Table 1 Kinetic parameters for cleavage of the P17 and P4 WT peptide substrates by
cathepsins S, L, B and D and by BACE1.
Enzymea
Substrateb
kcat
(s-1)
Km
(m
kcat/Km
M
Ratioc
Cat./BACE1
-1
M
)
s-1
"
"
"
19 200 2200
Cat.S-bov
Cat.S-bov
Cat.L-bov
Cat.B-bov
P17 (10)
P4 (14)
P17 (10)
P17 (10)
34.2 4.7
1.8 0.36
"
"
"
52.4 5.1
2.4 0.43
21 600 1100
"
13.8 0.8
"
3.7 0.15
"
3720 310
"
16 220 1500
"
"
"
Cat.S-hum
Cat.S-hum
Cat.L-hum
Cat.B-hum
P17 (10)
P4 (14)
P17 (10)
P17 (10)
148 34
2.7 0.61
54 700 4800
1170
570
74
"
"
"
137 40
3.1 0.70
44 800 9500
"
9.7 1.6
"
2.8 0.52
"
3460 580
"
20 470 1900
440
Cat.D-hum
Cat.D-hum
P17 (10)
P4 (14)
29.5
134
0.6
1.7
"
46.6 6.7
BACE1
BACE1
P17 (10)
P4 (14)
"
78.5 16.5
aKinetic constants of cathepsins S, L and B determined in the present study are presented
"
as means SD values from four experiments. Data of cathepsin D and BACE1 are from
Schechter and Ziv (2008). The kcat/Km values for the cleavage of P17 and P4 by BACE1
are average of values from two laboratories (Schechter and Ziv, 2008). Cat, cathepsin.
bSubstrate designation (P17 or P4) and the number of amino acid residues in the substrate
(in brackets).
cRatio of kcat/Km for cleavage of the same WT peptide substrate by human cathepsin (S, L,
B or D) compared to human BACE1.
Human cathepsin S rapidly hydrolyzes the P17 (10-mer)
and P4 (14-mer) WT peptides with kcat/Km values of 54 700
would expose new N-termini: AspAla«, AlaGlu«, Glu-
Phe«, etc. (Figure 1). Aminopeptidase readily cleaves free
dipeptides, whereas carboxypeptidase and endopeptidases
weakly hydrolyze free dipeptides (Schechter and Berger,
1966b, 1967). Therefore, we used four dipeptides corre-
sponding to the above N-termini (MetAsp, AspAla, AlaGlu,
GluAla instead of GluPhe) as substrates for the cytosol and
microsomal aminopeptidases.
Cleavage of the dipeptides was measured at two enzyme
concentrations (2 mg/ml and 20 mg/ml) after 1 h and 10 h
incubation. Hydrolysis of dipeptides by 2 mg/ml enzyme
after 1 h and 10 h is shown (Figure 8). Both enzymes rapidly
cleave MetAsp: 10–25% after 1 h and approximately 90%
after 10 h. The AlaGlu dipeptide is cleaved somewhat slower
than MetAsp, and GluAla is cleaved similar to AlaGlu. On
the other hand, the AspAla dipeptide is a poor substrate.
Hydrolysis of AspAla was not detected after 1 h and 10 h
incubation with cytosol aminopeptidase. With microsomal
aminopeptidase, the AspAla was slightly hydrolyzed after
1 h (approx. 3%). After 10 h, hydrolysis increased but it was
significantly less than that observed with the MetAsp,
AlaGlu and GluAla dipeptides (Figure 8).
Table 2 gives the information that can be extracted from
Figure 8 (2 mg/ml enzyme) and from enzyme concentration
at 20 mg/ml, regarding the effect of the N-terminal residue
on susceptibility to hydrolysis by the two aminopeptidases.
The Asp residue at the N-terminus is fairly resistant to
hydrolysis. The cytosol aminopeptidase cleaves the N-ter-
minal Met, Ala and Glu residues more efficiently: 92-, 35-
and 44-fold faster than Asp. The microsomal enzyme cleaves
N-terminal Met, Ala and Glu residues 21-, 8.1- and 6.6-fold
-1
and 44 800
M
s-1, respectively. Human cathepsin B rapidly
hydrolyzes the P17 peptide but less rapidly than cathepsin S
-1
(kcat/Km: 20 470
M
s-1). The human cathepsin L hydrolyzes
-1
the P17 peptide significantly slower (kcat/Km: 3460
M
s-1)
than the human cathepsins S and B (Table 1). Human
BACE1 cleaves the P17 and P4 WT peptides very slowly
-1
with kcat/Km values of 46.6 and 78.5
M
s-1, respectively.
Human cathepsin D has comparably poor activity toward
these peptides (Schechter and Ziv, 2008). Thus, cathepsins
S, L and B cleave the WT peptides much faster than BACE1
and cathepsin D. Cathepsin S cleaves the P17 and P4 pep-
tides 1170- and 570-fold faster than BACE1, respectively.
Cathepsins B and L cleave the P17 peptide 440- and 74-fold
faster than BACE1, respectively (Table 1).
BACE1 and cathepsin D split the P17 and P4 peptides at
the WT b-site (M-D bond, Schechter and Ziv, 2008), where-
as cathepsins S, B and L cleave at two bonds flanking the
WT b-site (K-MD-A). Cleavage at the major K-M bond
(85%) would yield a modified Ab extended at the N-termi-
nus by one Met residue. Therefore, we studied aminopepti-
dases to evaluate if they would rapidly remove the
N-terminal Met and slowly remove N-terminal Asp, to favor
the formation of authentic Ab initiated with Asp1.
Aminopeptidases cleave N-terminal Met faster
than N-terminal Asp
The CqM and P6qM cleavage products have N-terminal
MetAsp«, and sequential degradation by aminopeptidase
Brought to you by | University of California - San Francisco
Authenticated
Download Date | 12/13/14 1:15 AM

Article in press - uncorrected proof
562 I. Schechter and E. Ziv
faster than Asp (Table 2, Figure 8). Both Asp and Glu are
negatively charged. However, the N-terminal Glu is suscep-
tible to hydrolysis whereas N-terminal Asp is weakly hydro-
lyzed (Figure 8, Table 2).
Rapid cleavage of N-terminal Met compared to Asp was
also observed with longer peptides derived from the N-ter-
minus of Ab, that is, the P6qM (7-mer) and P6 (6-mer)
peptides. Cytosol aminopeptidase (2 mg/ml) cleave the N-
terminal Met of P6qM to yield Met and P6 (approx. 15%
after 1 h, 100% after 10 h), and the P6 product is not further
hydrolyzed. In agreement, the P6 substrate with N-terminal
Asp is not cleaved after 1 h and weakly (approx. 3%) after
10 h (Figure 8). Based on these findings and data of cytosol
aminopeptidase at 20 mg/ml (data not shown), we estimate
that the N-terminal Met residue of P6qM is cleaved approx-
imately 40-fold faster than the N-terminal Asp residue of P6
(Figure 8, Table 2).
The microsomal aminopeptidase cleaves the P6qM and
P6 peptides faster than the cytosol enzyme (Figure 8). In-
spection of the cleavage products indicates that this enzyme
contains an additional endopeptidase activity that compli-
cates interpretation of the data. After 1 h incubation the
P6qM and P6 peptides reveal a spot with slow mobility,
which disappears after 10 h (Figure 8B). The same spot is
also observed when the 25V fraction of the brain extract was
reacted with P6qM and P6 (Figure 9). It is likely that this
spot is a peptide generated by endopeptidase activity and it
serves as substrate to aminopeptidase. Therefore, at this
stage, we did not evaluate the relative rates of N-terminal Met
or Asp release from P6qM or P6 by the microsomal enzyme.
It was now of interest to evaluate aminopeptidase activity
in the brain. The 25V and S100 fractions of bovine brain
extract were chromatographed on G-100 columns and ali-
quots of the eluate were reacted with the peptides. The results
showed reaction patterns similar to those observed with the
purified aminopeptidases. The MetAsp dipeptide was
degraded. The AlaGlu and GluAla dipeptides were degraded
slower or at comparable rate to MetAsp, whereas the AspAla
dipeptide was not degraded. The N-terminal Met residue of
P6qM was cleaved whereas the N-terminal Asp residue of
P6 was not (Figure 9).
Figure 8 Cytosol and microsomal aminopeptidases cleave N-ter-
minal Met faster than N-terminal Asp.
Cytosol (A1, A2) and microsomal (B1, B2) aminopeptidases
(2 mg/ml) were reacted with the P6qM and P6 peptides and with
the dipeptides MetAsp (MD), AspAla (DA), AlaGlu (AE) and
GluAla (EA) for 1 h (A1, B1) and 10 h (A2, B2). The positions of
amino acid markers Arg (R), Gly (G), Ala (A), Met (M), Glu (E),
Phe (F) and Asp (D), in addition to ArgGly (RG), P6qM and P6
are shown on the left. Asterisk in B1 indicates a peptide likely
formed by endopeptidase activity contaminating the microsomal
aminopeptidase. Samples were resolved by paper electrophoresis in
0.5
M
formic acid.
Table 2 Aminopeptidases cleave N-terminal Met
faster than N-terminal Asp.
The 25V fraction reveals two additional peptides from
P6qM and P6: one with slow mobility and one moving rap-
idly corresponding to mobility of the ArgGly dipeptide (Fig-
ure 9). This is probably due to endopeptidase activity. We
speculate that it cleaves the F-R bond to yield the Met-
AspAlaGluPhe pentapeptide from P6qM or the AspAla-
GluPhe tetrapeptide from P6 plus the dipeptide ArgGly
(Figure 1). The ArgGly dipeptide is then cleaved into argi-
nine and glycine (Figure 8B). The MetAspAlaGluPhe is rap-
idly cleaved into Met plus AspAlaGluPhe. We propose that
the peptide indicated by an asterisk (Figures 8 and 9) is the
presumed AspAlaGluPhe tetrapeptide. Indeed, its electropho-
Peptide pair
Pep1/Pep2 ratio of rate
of cleavage by
aminopeptidasea
Cytosol
Microsomal
P6qM (Met)/P6 (Asp)b
MetAsp/AspAla
AlaGlu/AspAla
GluAla/AspAla
aPep1/Pep2 ratio of rate of cleavage rate is average of
two experiments.
40
4
8
n.d.
"
21 4
"
"
92
"
"
44
"
"
6.6 1.4
35
5
4
8.1 1.9
bCleavage of N-terminal Met of P6qM compared to
cleavage of N-terminal Asp of P6 by cytosol amino-
peptidase. Ratio for cleavage of the same peptides by
microsomal aminopeptidase could not be determined
(n.d.) due to possible contamination by endopeptidase
(see Figure 8 and text).
retic mobility at 0.5
M formic acid corresponds to the expect-
ed mobility for this acidic tetrapeptide (Schechter and
Berger, 1966a).
The aminopeptidase activity overlaps and flanks the void
volume of the G-100 column (G150 000 kDa) and it is
Brought to you by | University of California - San Francisco
Authenticated
Download Date | 12/13/14 1:15 AM

Article in press - uncorrected proof
New b-secretase candidates rapidly cleave the WT b-site 563
Figure 9 Aminopeptidase activity in the brain cleaves N-terminal Met faster than N-terminal Asp.
The 25V (left column) and S100 (right column) fractions of bovine brain were loaded on a Sephadex G-100 column. Aminopeptidase
activity was tested by mixing aliquots of the eluate with the P6qM and P6 peptides and with several dipeptides. Positions of the peptide
substrates (bold), of the cleavage products ArgGly (RG) and of amino acids (single letter code) are shown on the left. Asterisk indicates a
peptide probably formed by endopeptidase activity of the 25V fraction eluted near the void volume. The positions of two size markers are
shown: dextran blue (DXB, indicates the void volume G150 kDa) and BSA (67 kDa). Samples were resolved by paper electrophoresis in
0.5
M formic acid.
slower than the BSA marker (67 kDa, Figure 9). This posi-
tion conforms with the size of aminopeptidases that form
large aggregates that range in mass from 100 to 400 kDa
(Taylor, 1993). In addition, the active fractions cleave the
AlaAla dipeptide into alanine, the AlaAla-NH2 is split into
alanine plus Ala-NH2, and N-acetyl-AlaAla is not cleaved,
as expected for an aminopeptidase (data not shown). Addi-
tional aminopeptidase activity was not detected in G-100
fractions down to the position of the 14 kDa lysozyme mark-
er (data not shown).
Cathepsin B is approximately 75-fold more abundant
than BACE1 in human brain
Because cathepsins S, B and L cleave the WT b-site
sequence, it was of interest to estimate the amount of these
enzymes in human brain. Quantitative Western blots were
performed by using increasing amounts of human cathepsins
as calibration standard (1–20 ng) and increasing amounts of
total protein extracts of cortex and hippocampus obtained from
human brain (2–8 mg) as described (Schechter and Ziv, 2008).
Brought to you by | University of California - San Francisco
Authenticated
Download Date | 12/13/14 1:15 AM

Article in press - uncorrected proof
564 I. Schechter and E. Ziv
Antibodies to cathepsin B detect cathepsin B, but not
BACE1 or cathepsins S, L and D. Scanning of the blots
shows linear dependence of densitometry units to the amount
of cathepsin B or brain protein loaded on the gel. From these
data we estimate that the amount of cathepsin B (ng enzyme/
cleaved) where P2 is a hydrophobic amino acid residue
(Schechter and Berger, 1968; Schechter and Ziv, 2006). In
agreement, the two susceptible bonds in the APP WT b-site
sequence are preceded by a hydrophobic P2: Val (VK-M) or
Met (MD-A). Cleavage of the K-M bond in the WT b-site
sequence was previously reported. Cells transfected by
cathepsin S cleave at the K-M bond to generate Ab peptide
extended at the N-terminus by one Met residue (Munger et
al., 1995). Peptides spanning the WT b-site sequence are
cleaved at the K-M bond by cathepsin B (Bohme et al., 2008)
and cathepsin L (Sahasrabudhe et al., 1993). Neuroblastoma
cells transfected by human WT or SW APP secrete into the
culture medium Ab peptides generated by cleavage at the
D-A bond (Wang et al., 1996).
"
mg brain protein) is 1.6 0.3 ng/mg in the cortex and
"
1.4 0.3 ng/mg in the hippocampus of human brain (average
values from two experiments). Others reported 0.02 ng
BACE1/mg protein in human cortex (Yang et al., 2003).
Thus, in human brain cathepsin B is approximately 75-fold
(1.5/0.02) more abundant than BACE1.
We failed to detect cathepsin L in brain extracts because
antibodies to cathepsin L (three commercial preparations) did
not detect the control of purified cathepsin L loaded on the
gel. However, immunohistochemistry reveals that cathepsin
L is abundant in the neurons of normal human brain and in
the neurons and amyloid plaques of Alzheimer’s patients,
similar to cathepsin B (Cataldo and Nixon, 1990; Cataldo et
al., 1991).
Antibodies to cathepsin S specifically detect cathepsin S
(1–20 ng) but not BACE1 and not cathepsins B, L and D.
However, specific immunoreactive bands were not observed
in the brain extracts, likely due to low concentration of the
enzyme in normal human brain. It was previously shown that
cathepsin S is weakly detected in a small minority of normal
human brain, but it is upregulated and expressed in the brain
of Alzheimer’s patients (Lemere et al., 1995; Munger et al.,
1995).
The interest in AD has led to the study of many enzymes
as b-secretase candidates (Hussain et al., 1999). To our
knowledge, except for BACE1 and recently cathepsin B,
these enzymes have not been further investigated. This report
deals with properties of cathepsins S, B and L in relation to
Ab-peptide formation that were not known before, these
include: kinetic constants, two cleavage points at the WT b-
site sequence and the role of aminopeptidases to produce the
Ab-peptide. Thus, it was reported that transfection of cells
with cathepsin S increases the secretion of modified Ab (N-
terminal Met) into the culture medium, suggesting that
cathepsin S could be relevant to the pathogenesis of AD
(Munger et al., 1995). However, kinetic parametrs for cleav-
age at the WT b-site sequence were not determined, two
cleavage points at the WT b-site sequence were not identi-
fied, and the need for aminopeptidase to generate the authen-
tic Ab (N-terminal Asp) was not shown. One research group
identified cathepsin B as the b-secretase (Hook et al., 2005)
and another group evaluated properties of the enzyme in this
context (Bohme et al., 2008). The published data on cathep-
sin B differ on points of cleavage in the WT b-site sequence,
that is, cleavage at the M-D bond (Hook et al., 2008a), main-
ly at the K-M bond, or at multiple bonds (Bohme et al.,
2008). The kcat/Km values for cleavage of the WT b-site pep-
Discussion
Because BACE1 has very low activity toward the WT b-site
(kcat/Km: approx. 50 M
-1 s-1) we searched for other b-secretase
candidates and identified three cathepsins that split the WT
b-site sequence better than BACE1. Human cathepsin S
-1
s-1) 1170-fold
tide substrates differ extensively, that is, 317 000
M
-1 s-1 (Hook
cleaves the WT b-site (kcat/Km: 54 700
M
-1
et al., 2008a) or 458
M
s-1 (Bohme et al., 2008). It is likely
better than BACE1, whereas cathepsin B (kcat/Km: 20 470
-1
-1
M
s-1) and cathepsin L (kcat/Km: 3460
M
s-1) are 440- and
that the above discrepancies are due to the use of peptide
substrates bearing markedly hydrophobic groups (methylcou-
marinamide; 5-w(2-amidoethyl) aminox naphthalene-1-sulfo-
nyl acid; 4-(4-dimethylaminophenyl-azo)benzoyl) in close
proximity to the b-site sequence, that probably affects the
rate of hydrolysis and point of cleavage. We have used the
P17 peptide substrate that spans a larger portion of the WT
b-site sequence, without added hydrophobic groups. This
WT b-site peptide is cleaved by cathepsin B with kcat/Km of
74-fold better than BACE1, respectively (Table 1). Western
blots of brain extracts reveal that cathepsin B is 75-fold more
abundant than BACE1. We failed to detect cathepsin L by
Western blotting but earlier immunohistochemical studies
showed that cathepsin L levels in the brain are similar to that
of cathepsin B (Cataldo et al., 1991) and that cathepsins B,
L and D are the most abundant lysosomal proteases (Turk et
al., 2000). The expression of cathepsin S in normal brain is
very low but it is induced in the brain of AD patients
(Lemere et al., 1995).
The three cathepsins split the WT sequence at two bonds
flanking the WT b-site (K-MD-A) where the K-M bond
(85%) is cleaved more frequently than the D-A bond (15%).
Cathepsins S, B and L are cysteine proteases of the papain
family (Rawlings and Barrett, 1993) in which specificity is
determined by a hydrophobic subsite S2, that is, cleavage
occurs next to P2 (P2P1-P19, hyphen indicates the bond
20 470
M
-1 s-1 and analyses of the cleavage products enabled
unambiguous determination of the bonds split, that is, 85%
at the K-M bond, 15% at the D-A bond and none at the
M-D bond. Cathepsin L was not implicated in AD.
Cleavage at the major K-M bond yields modified Ab
extended at the N-terminus by one Met residue that should
be removed to generate the authentic Ab initiated by Asp1.
To address this issue, we examined the activity of cytosol
and microsomal aminopeptidases on peptides corresponding
Brought to you by | University of California - San Francisco
Authenticated
Download Date | 12/13/14 1:15 AM

Article in press - uncorrected proof
New b-secretase candidates rapidly cleave the WT b-site 565
to Ab extended with N-terminal Met, the regular Ab with
N-terminal Asp and dipeptides corresponding to N-termini
of intermediate cleavage products generated by sequential
degradation of Ab by aminopeptidases. It was found that the
enzymes rapidly remove the N-terminal Met but not the N-
terminal Asp, which is cleaved 21- to 92-fold slower than
Met (Figure 8, Table 2). Aminopeptidase activity in the brain
reveals similar specificity, that is, N-terminal Met is readily
hydrolyzed whereas N-terminal Asp is weakly hydrolyzed
(Figure 9). The major cathepsin cleavage product of Ab
extended by N-terminal Met has not been detected in the
brain of AD patients (Munger et al., 1995). This is probably
due to the specificity of aminopeptidases that rapidly convert
the modified Ab (N-terminal Met) into regular Ab peptide
(N-terminal Asp). Poor susceptibility of Asp to hydrolysis
by aminopeptidases would prolong the in vivo lifetime of
regular Ab that aggregates to form amyloid plaques.
Brain aminopeptidases were classified to a major fraction
()80%) of Ala-aminopeptidase that cleaves neutral and
basic amino acids but not N-terminal acidic residues and a
minor fraction of Asp-aminopeptidase that cleaves N-termi-
nal acidic amino acids (Asp, Glu) (Iribar et al., 1995). It was
reported that the brain Asp-aminopeptidase is selectively
decreased in aged brain (Iribar et al., 1995), a process that
might reduce Ab catabolism, prolong Ab retention in the
aging brain, and thus favor its deposition as plaques (Saido,
1998). Trimming by exopeptidases is a common process in
the brain involving neuropeptide hormone maturation (Hook
et al., 1994). In other tissues the precursors of several cathep-
sins (D, B, L, H) are processed by endopeptidases and are
then trimmed by aminopeptidases to produce the mature
active protease (Erickson, 1989).
some and endosome vesicles but they were also found in
vivo in the extracellular space, on the cell membrane, in the
cytosol and inside nuclei (Turk et al., 2000; Tedelind et al.,
2010). The cysteine cathepsins B and L and the aspartic
cathepsin D are ubiquitous and they are the most abundant
lysosomal proteases (Turk et al., 2000). Cathepsins S and K
exhibit restricted and regulated tissue expression (Chapman
et al., 1997). The presenilin complex was initially found in
the endoplasmic reticulum and Golgi. However, it was later
shown that Presenilin-1, Nicastrin, APP and g-secretase
activity colocalize in the lysosomal membrane (Pasternak et
al., 2003).
High levels of cathepsins B and L were found in neurons
and amyloid plaques in the brain of AD patients (Cataldo et
al., 1991). Studies on brains from early stages of Down syn-
drome and of AD patients reveal the appearance of activated
or enlarged endosomes containing soluble Ab prior to amy-
loid deposition (Cataldo et al., 2004). Cathepsin S is weakly
detected in normal human brain, but in AD brain the enzyme
is readily detected (Lemere et al., 1995).
Knockout of the cathepsin B gene reduces the levels of
Ab and of the C-terminal b-secretase fragment (CTFb) in
the brain of mice expressing the human WT APP (Hook et
al., 2009). The cysteine protease inhibitor E64d and the relat-
ed inhibitor CAO74Me (preferentially inhibits cathepsin B)
infused into the brain of a mouse AD model reduced amyloid
plaques, decreased Ab and CTFb levels in the brain and
improved memory deficit of the animals (Hook et al.,
2008a). Transfection of human kidney cells by cathepsin S
increased the secretion of Ab and E-64d reduced Ab secre-
tion by the cathepsin S transfected cells (Munger et al.,
1995).
Further studies in cells and animals are needed to establish
that cathepsins S, B and L are relevant in vivo to the for-
mation of the Ab peptide and to the pathogenesis of AD.
However, studies on the in vivo activity of these enzymes in
relation to AD have been published. Summarized below is
information on the colocalization of cathepsins and site of
Ab formation, the effect of cysteine protease inhibitors on
the level of Ab in cells and in AD animal models, and
knockout of cathepsin B in APP transgenic mice.
The localization of APP processing is controversial. How-
ever, many studies implicate the endosome and lysosome
compartment as the major site of Ab formation. APP con-
tains the re-internalization signal and APP molecules from
the cell surface are targeted intracellularly to the endosome
and lysosome vesicles in which APP amyloidogenic frag-
ments have been found (Golde et al., 1992; Haass et al.,
1992). This acidic compartment contains cathepsins that
potentially could process the amyloidogenic fragments to
form the Ab peptide. In the past 20 years it was realized that
proteolytic enzymes are key players in a wide range of bio-
logical processes in health and disease (Turk et al., 2000;
Hooper, 2002). Cathepsins are involved in bone remodeling
and antigen presentation (Chapman et al., 1997), tumor
growth and metastasis (Mohamed and Sloane, 2006), thy-
roglobulin processing (Friedrichs et al., 2003) and other
processes. The cathepsins are mainly localized in the lyso-
Because the specificity of cathepsins S, B and L is deter-
mined mainly by a hydrophobic S2, peptide with penultimate
hydrophobic P2 are inhibitors due to the interaction with the
subsite S2 (Schechter and Berger, 1968; Schechter, 2005). It
was reported that N-blocked di- and tri- peptide aldehydes
with penultimate hydrophobic P2 inhibit the secretion of Ab
from cells transfected with APP in a dose dependent manner
(Sinha and Lieberburg, 1999). The tripeptide aldehyde
Z.ValLeuLeu-CHO inhibited the secretion of Ab and anal-
ysis of the cellular intermediates of APP degradation prod-
ucts indicated inhibition of the b-secretase (Abbenante et al.,
2000). Infusion of Ac.LeuValLys-CHO into the brain of
guinea pigs decreased Ab in the brain (Hook et al., 2007).
In summary, cathepsins S, B and L with aminopeptidases
display b-secretase activity. These cathepsins cleave the WT
b-site sequence much better than BACE1; cathepsins B and
L are more abundant than BACE1; the expression of cathep-
sin S is induced in the brain of AD patients; these cathepsins
are localized in the endosome and lyzosome system where
Ab formation likely occurs; cysteine protease inhibitors can
reduce the production of Ab in cells and animals, in addition
to improving memory deficit of the AD animal model. Taken
together, these findings indicate that cathepsins S, B and L
are promising therapeutic targets to develop inhibitors serv-
ing as disease modifying drugs that would slow down, pre-
vent or reverse the progress of Alzheimer’s disease.
Brought to you by | University of California - San Francisco
Authenticated
Download Date | 12/13/14 1:15 AM

Article in press - uncorrected proof
566 I. Schechter and E. Ziv
SDS-PAGE as described (Schechter and Ziv, 2008). The gel was
stained with Coomassie Blue, protein bands were excised and
digested by trypsin. Tryptic peptides were separated by reverse
phase HPLC (Model SP8800, Spectra Physics, San Jose, CA, USA).
Eluted peptides were directed online into a nanospray injector and
peptides were characterized by Q-STAR TOF tandem mass spec-
trometer (MDS-Sciex, Foster City, CA, USA).
This paper supports and extends earlier reports suggesting
that aminopeptidases and cathepsin S might play a role in
the pathogenesis of AD, that cathepsin B can function as b-
secretase and it does not imply that BACE1 is not involved
in the processing of APP in vivo. Processing of the WT APP
present in the worldwide population could be due to the joint
action of BACE1, cathepsins S, B and L, and aminopepti-
dases. Further research would clarify the relative contribution
of each enzyme at early and late stages of development of
the disease.
Many studies demonstrate that the presenilin complex is
associated with g-secretase activity. However, the role of the
presenilin complex in APP cleavage is not conclusive: direct
action as protease or indirect action as regulator of g-secre-
tase. Therefore, we searched and identified additional g-
secretase candidates. Early studies show that the new
enzymes cleave the g, z and ´ bonds in CTFb probably
faster than the presenilin complex. They are susceptible to
inhibition and modulation by a subset of non-steroidal anti-
inflammatory drugs more than the presenilin complex and
they are not associated with Notch. We suggest that enzy-
matic cleavage of CTFb is initiated at the ´-site located close
to the membrane cytosol interface, followed by sequential
exo/endo peptidase activity through the z-site until the g-
site, to form the neurotoxic Ab-peptide.
Peptide substrates and cleavage products
Two WT peptide substrates were used to assay b-secretase activity:
SEVKMDAEFR (P17, 10-mer) and EEISEVKMDAEFRG (P4, 14-
mer). Peptides produced by enzymatic cleavage of the WT P17 and
P4 substrates at the b-site (M-D bond) and at flanking sites (K-M
and D-A bonds) are shown in Figure 1. The WT P17 and P4 peptide
substrates were prepared in the department of Organic Chemistry
(Schechter and Ziv, 2008). All other peptides (cleavage products)
were synthesized by GL Biochem (Shanghai, China). The dipeptides
MetAsp, AspAla, AlaGlu and GluAla were purchased from Bachem
(Bubenorf, Switzerland).
b-Secretase activity at different purification stages
This was assayed in 10 ml reaction mixture containing: 2–5 ml of
test sample, 2 mM peptide substrate, 25 mM Na-acetate buffer pH
5.2, 10 m mercaptoethanol. After incubation (4–6 h at 378C) sam-
M
ples (3–6 ml) were applied on a Whatman number three paper and
components of the reaction mixture were resolved by paper electro-
phoresis run in the pH 6 buffer (2.9 mM citric acid/140 mM
Na2HPO4) or in 0.5 formic acid, and peptide spots were devel-
M
oped by ninhydrin, essentially as described (Schechter and Berger,
1966a). Enzyme digests of all peptides were analyzed by reverse
phase HPLC performed with a RP18 5 m column (Merck, Darm-
stadt, Germany) developed with acetonitrile gradient in 0.1% tri-
fluoroacetic acid (TFA) as indicated in the Figure legends. Elution
was monitored at 220 nm. Peptides in HPLC peaks were character-
ized by paper electrophoresis compared to synthetic peptide markers
and by MALDI-TOF mass spectrometry (reflex 3; Bruker, Bremen,
Germany).
Materials and methods
Purification of a putative b-secretase from bovine
kidney
Bovine kidney (2 h after sacrifice) was homogenized in 0.25
sucrose, 1 m DTT and buffer (SDB) (1 g tissue per 3 ml SDB
solution). Buffer solution was 20 m Tris HCl pH 7.3 or 20 m
M
M
M
M
Na-acetate pH 5.2. The homogenate was spun (3000 g, 10 min) to
sediment nuclei and unbroken tissue. The supernatant was spun
(25 000 g, 30 min). The pellet was suspended in SDB to yield the
25V fraction (vesicles sedimented at 25 000 g) enriched with endo-
some and lysosome (Schechter and Ziv, 2008). The supernatant was
spun again (100 000 g, 1 h) to yield the S100 fraction. The 25V
and S100 fractions were kept in liquid nitrogen until used. Triton
X-100 (1.5 mg detergent/1 mg protein) was added to the 25V frac-
tion in order to disrupt the 25V vesicles. After 20 min on ice the
solution was spun (8000 g, 10 min). The clear supernatant was
Determination of the kinetic constants kcat, Km
and kcat/Km
Human cathepsin S (recombinant, 24.5 kDa), cathepsin L (liver,
29 kDa) and cathepsin B (liver, 27.5 kDa) in addition to bovine
cathepsin S (spleen, 24 kDa), cathepsin L (kidney, 31 kDa) and
cathepsin B (spleen, 28 kDa) were purchased from Calbiochem
(Darmstadt, Germany). Enzyme concentration and molecular mass
were based on values provided by the manufacturer. Cathepsin
activity was determined in a reaction mixture (20–40 ml) containing
loaded on a Sephadex G-100 column (40 m Na-acetate pH 5.2,
M
1 m DTT or 40 m Tris HCl pH 7.3, 1 mM DTT). Aliquots of
M
M
1.8 m
sin S (12 m
L (14 m
M
DTT, 1.8 m
citric acid/43 m
citric acid/39 m Na2HPO4) and at pH 4.5 for cathepsin
citric acid/31 Na2HPO4). Duplicate aliquots
M
EDTA and 50 mg/ml BSA at pH 6 for cathep-
the eluate were analyzed for b-secretase activity by paper electro-
phoresis (see below) and proteins by SDS-PAGE. Active fractions
were pooled, concentrated by vivaspin concentrator (Sartorius, Bin-
brook Hill, UK) and loaded on CM52 cation exchanger equilibrated
M
M
Na2HPO4), at pH 5.5 for cathepsin
M
M
B
(18
m
M
mM
(6–15 ml) were removed from the reaction mixture after 6–20 min
incubation and from substrate hydrolyzed to completion (excess
enzyme and prolonged incubation). Components of the reaction
mixture were resolved by paper electrophoresis in the pH 6 buffer.
The paper sheet was developed by ninhydrin (Figure 2), the colored
spots were cut out, dipped in elution solution, centrifuged and the
color of the clear supernatant was quantified by measuring absor-
bance at 570 nm (Schechter and Berger, 1966a). Extent of substrate
hydrolysis was determined from the color of cleavage products in
the test sample and the color of the same cleavage products obtained
with 40 mM Na-acetate pH 5.4, 1 mM DTT, 0.05% Triton X-100.
The b-secretase activity was found in the effluent that was collected,
concentrated and further purified by electrophoresis in neutral (pH
7.2) and acidic (pH 5.4) polyacrylamide gels (Maizel, 1971). These
gels provide resolution based on the distinct pI of the protein, caus-
ing different electrophoretic mobility at different pH values. Fur-
thermore, b-secretase activity was recovered from the gels in good
yields (G75% of activity loaded on the gel). Active fractions from
the acidic gel were resolved on 8–14% polyacrylamide gradient
Brought to you by | University of California - San Francisco
Authenticated
Download Date | 12/13/14 1:15 AM

Article in press - uncorrected proof
New b-secretase candidates rapidly cleave the WT b-site 567
from substrate hydrolyzed to completion, as described (Schechter
and Ziv, 2008). Kinetic constants of cathepsins S and L were derived
from Lineweaver-Burk plots of P17 and P4 degradation at nine dif-
ferent concentrations (0.52–2.4 mM). For cathepsin B only the kcat/Km
constant was determined (Results section).
Cataldo, A.M., Petanceska, S., Terio N.B., Peterhoff, C.M., Durham,
R., Mercken, M., Mehta, P.D., Buxbaum, J., Haroutunian, V. and
Nixon, R.A. (2004). A b localization in abnormal endosomes:
association with earliest Ab elevations in AD and Down syn-
drome. Neurobiol. Aging 25, 1263–272.
Chapman, H.A., Riese, R.J., and Shi, G.P. (1997). Emerging roles
for cysteine proteases in human biology. Annu. Rev. Physiol. 59,
63–88.
Aminopeptidase activity assays
Cole, S.L. and Vassar, R. (2008). The role of amyloid precursor
protein processing by BACE1, the b-secretase, in Alzheimer dis-
ease pathophysiology. J. Biol. Chem. 283, 29621–29625.
Dunn, B.M. and Hung, S. (2000). The two sides of enzyme-sub-
strate specificity: lessons from the aspartic proteinases. Biochim.
Biophys. Acta 1477, 231–240.
Erickson, A.H. (1989). Biosynthesis of lysosomal endopeptidases.
J. Cell. Biochem. 40, 31–41.
Friedrichs, B., Tepel, C., Reinheckel, T., Deussing, J., von Figura,
K., Herzog, V., Peters, C, Saftig, P., and Brix, K. (2003). Thyroid
functions of mouse cathepsins B, K, and L. J. Clin. Invest. 111,
1733–1745.
Golde, T.E., Estus, S., Younkin, L.H., Selkoe, D.J., and Younkin,
S.G. (1992). Processing of the amyloid protein precursor to
potentially amyloidogenic derivatives. Science 255, 728–730.
Haass, C., Koo, E.H., Mellon, A., Hung, A.Y., and Selkoe, D.J.
(1992). Targeting of cell-surface b-amyloid precursor protein to
lysosomes: alternative processing into amyloid-bearing frag-
ments. Nature 357, 500–503.
Hook, V.Y., Azaryan, A.V., Hwang, S.R., and Tezapsidis, N. (1994).
Proteases and the emerging role of protease inhibitors in pro-
hormone processing. FASEB J. 8, 1269–1278.
Hook, V., Toneff, T., Bogyo, M., Greenbaum, D., Medzihradszky,
K.F., Neveu, J., Lana, W., Hook, G., and Reisine, T. (2005).
Inhibition of cathepsin B reduces b-amyloid production in reg-
ulated secretory vesicles of neuronal chromaffin cells: evidence
for cathepsin B as a candidate b-secretase of Alzheimer’s dis-
ease. Biol. Chem. 386, 931–940.
Cytosol leucine aminopeptidase (L5658, Sigma, St. Louis, MO,
USA) and microsomal leucine aminopeptidase (L5006, Sigma) were
used. Enzyme concentration was based on values provided by the
manufacturer. The cytosol aminopeptidase reaction mixture con-
tained enzyme, peptide (3 m
MgSO4 and 100 mg/ml BSA. The microsomal aminopeptidase reac-
tion mixture contained enzyme, peptide (3 m ), 25 m Tris HCl
pH 7.2, 2.5 m MgSO4 and 100 mg/ml BSA. The 25V and S100
fractions of bovine brain were chromatographed on a G100 column
equilibrated with 40 m Tris HCl pH 7.3 and 1 m DTT. Aliquots
from the effluent were reacted with peptide (3 m ) in presence of
5 m MgSO4. Reaction mixtures were resolved by paper electro-
phoresis in 0.5 formic acid.
M), 50 mM Tris HCl pH 8, 5 mM
M
M
M
M
M
M
M
M
Western blotting
Total protein extracts of human brain cerebral cortex and hippocam-
pus (Clontech, Palo Alto, CA, USA), goat antibodies to the enzymes
(Santa Cruz Biotechnology, Santa Cruz, CA, USA), and horse radish
peroxidase anti-goat antibodies (Jackson Immune Research Labo-
ratories, West Grove, PA, USA) were purchased from commercial
sources. Protein extracts and enzymes were resolved by SDS-PAGE,
electroblotted onto PVDF membranes and the relevant enzymes
were immunodetected as described (Schechter and Ziv, 2008).
References
Hook, G., Hook, V.Y., and Kindy, M. (2007). Cysteine protease
inhibitors reduce brain b-secretase activity in vivo and are poten-
tial Alzheimer’s disease therapeutics. Biol. Chem. 388, 979–983.
Hook, V., Kindy, M., and Hook. G. (2008a) Inhibitors of cathepsin
B improve memory and reduce b-amyloid in transgenic Alzhei-
mer disease mice expressing the wild-type, but not the Swedish
mutant, b-secretase site of the amyloid precursor protein. J. Biol.
Chem. 283, 7745–7753.
Abbenante, G., Kovacs, D.M., Leung, D.L., Craik, D,J., Tanzi, R.E.,
and Fairlie, D.P. (2000). Inhibitors of b-amyloid formation based
on the b-secretase cleavage site. Biochem. Biophys. Res.
Commun. 268, 133–135.
Aronson, N.N. and Barrett, A.J. (1978). The specificity of cathepsin
B. Hydrolysis of glucagon at the C-terminus by a peptidyldi-
peptidase mechanism. Biochem. J. 171, 759–765.
Bohme, L., Hoffmann, T., Manhart, S., Wolf, R., and Demuth, H.U.
(2008). Isoaspartate-containing amyloid precursor protein-
derived peptides alter efficacy and specificity of potential b-
secretases. Biol. Chem. 389, 1055–1066.
Hook, V., Schechter, I., Demuth, H.U., and Hook, G. (2008b). Alter-
native pathways for production of b-amyloid peptides of Alz-
heimer’s disease. Biol. Chem. 389, 993–1006.
Hook, V.Y., Kindy, M., Reinheckel, T., Peters, C., and Hook, G.
(2009). Genetic cathepsin B deficiency reduces b-amyloid in
transgenic mice expressing human wild-type amyloid precursor
protein. Biochem. Biophys. Res. Commun. 386, 284–288.
Hooper, M.N. (2002). Proteases in Biology and Medicine: Essays
in Biochemistry, Vol. 38 (Oxford, UK: Portland Press).
Hussain, I., Powell, D., Howlett, D.R., Tew, D.G., Meek, T.D.,
Chapman, C., Gloger, I.S., Murphy, K.E., Southan, C.D., Ryan,
D.M., et al. (1999). Identification of a novel aspartic protease
(Asp 2) as b-secretase. Mol. Cell Neurosci. 14, 419–427.
Iribar, C., Esteban, M.J., Martinez, J.M., and Peinado, J.M. (1995).
Decrease in cytosolic aspartyl-aminopeptidase but not in alanyl-
aminopeptidase activity in the frontal cortex of the aged rat.
Brain Res. 687, 211–213.
Bromme, D., Bonneau, P.R., Lachance, P., Wiederanders, B.,
Kirschke, H., Peters, C., Thomas, D.Y., Storer, A.C., and Vernet,
T. (1993). Functional expression of human cathepsin S in Sac-
charomyces cerevisiae. Purification and characterization of the
recombinant enzyme. J. Biol. Chem. 468, 4832–4880.
Cai, H., Wang, Y., McCarthy, D., Wen, H., Borchelt, D.R., Price,
D.L., and Wong, P.C. (2001). BACE1 is the major b-secretase
for generation of Abeta peptides by neurons. Nat. Neurosci. 4,
233–234.
Cataldo, A.M. and Nixon, R.A. (1990). Enzymatically active lyso-
somal proteases are associated with amyloid deposits in Alzhei-
mer brain. Proc. Natl. Acad. Sci. USA 87, 3861–3865.
Cataldo, A.M., Paskevich, P.A., Kominami, E., and Nixon, R.A.
(1991). Lysosomal hydrolases of different classes are abnormal-
ly distributed in brains of patients with Alzheimer disease. Proc.
Natl. Acad. Sci. USA 88, 10998–11002.
Jacks, T. (1996). Tumor suppressor gene mutations in mice. Annu.
Rev. Genet. 3, 603–636.
Brought to you by | University of California - San Francisco
Authenticated
Download Date | 12/13/14 1:15 AM

Article in press - uncorrected proof
568 I. Schechter and E. Ziv
Lemere, C.A., Munger, J.S., Shi, G.P., Natkin, L., Haass, C., Chap-
man, H.A., and Selkoe, D.J. (1995). The lysosomal cysteine pro-
tease, cathepsin S, is increased in Alzheimer’s disease and Down
syndrome brain. An immunocytochemical study. Am. J. Pathol.
146, 848–860.
Lin, X., Koelsch, G., Wu, S., Downs, D., Dashti, A., and Tang, J.
(2000). Human aspartic protease memapsin 2 cleaves the b-
secretase site of b-amyloid precursor protein. Proc. Natl. Acad.
Sci. USA 97, 1456–1460.
Lou, Y., Bolon, B., Kahn, S., Bennett, B.D., Babu-Khan, S., Denis,
P., Fan, W., Kha, H., Zhang, J., Gong, Y., et al. (2001). Mice
deficient in BACE1, the Alzheimer’s b-secretase, have normal
phenotype and abolished b-amyloid generation. Nat. Neurosci.
4, 231–232.
Maizel, J.V. (1971). Polyacrylamide gel electrophoresis of viral pro-
teins. Methods Virol. 5, 179–216.
Mohamed, M.M. and Sloane, B.F. (2006). Cysteine cathepsins: mul-
tifunctional enzymes in cancer. Nat. Rev. Cancer. 6, 764–
775.
Munger, J.S., Haass, C., Lemere, C.A., Shi, G.P., Wong., W.S.,
Teplow, D.B., Selkoe, D.J., and Chapman, H.A. (1995). Lyso-
somal processing of amyloid precursor protein to Ab peptides:
a distinct role for cathepsin S. Biochem. J. 311, 299–305.
Neve, R.L., Finch, E.A., and Dawes, L.R. (1988). Expression of the
Alzheimer amyloid precursor gene transcripts in the human
brain. Neuron 1, 669–677.
Pasternak, S.H., Bagshaw, R.D., Guiral, M., Zhang, S., Ackerley,
C.A., Pak, B.J., Callahan, J.W., Mahuran, D.J. (2003). Preseni-
lin-1, nicastrin, amyloid precursor protein, and gamma-secretase
activity are co-localized in the lysosomal membrane. J. Biol.
Chem. 278, 26687–26694.
Schechter, I. and Ziv, E. (2006). Inhibitors can activate proteases to
catalyze the synthesis and hydrolysis of peptides. Biochemistry
45, 14567–14572.
Schechter, I. and Ziv, E. (2008). Kinetic properties of cathepsin D
and BACE 1 indicate the need to search for additional b-secre-
tase candidate(s). Biol. Chem. 389, 313–320.
Schilling, S., Lauber, T., Schaupp, M., Manhart, S., Scheel, E.,
Bo¨hm, G., and Demuth, H.U. (2006). On the seeding and olig-
omerization of pGlu-amyloid peptides (in vitro). Biochemistry
45, 12393–12399.
Schilling, S., Zeitschel, U., Hoffmann, T., Heiser, U., Francke, M.,
Kehlen, A., Holze, R.M., Hutter-Paier, B., Prokesch, M., Win-
disch, M., et al. (2008). Glutaminyl cyclase inhibition attenuates
pyroglutamate Ab and Alzheimer’s disease-like pathology. Nat.
Med. 14, 1106–1111.
Selkoe, D.J. (2001). Alzheimer’s disease: genes, proteins, and ther-
apy. Physiol. Rev. 81, 741–766.
Shi, X.P., Chen, E., Yin, K.C., Na, S., Garsky, V.M., Lai, MT., Li,
Y.M., Platchek, M., Register, R.B., Sardana, M.K., et al. (2001).
The pro domain of b-secretase does not confer strict zymogen-
like properties but does assist proper folding of the protease
domain. J. Biol. Chem. 276, 10366–10373.
Shi, X.P., Tugusheva, K., Bruce, J.E., Lucka, A., Chen-Dodson, E.,
Hu, B., Wu, G.X, Price, E., Register, R.B., Lineberger, J., et al.
(2005). Novel mutations introduced at the b-site of amyloid b
protein precursor enhance the production of amyloid beta pep-
tide by BACE1 in vitro and in cells. J. Alzheimer’s Dis. 7,
139–148.
Sinha, S. and Lieberburg, I. (1999). Cellular mechanisms of b-amy-
loid production and secretion. Proc. Natl. Acad. Sci. USA 96,
11049–11053.
Rawlings, N.D. and Barrett, A.J. (1993). Evolutionary families of
peptidases. Biochem. J. 290, 205–218.
Sinha, S., Anderson, J.P., Barbour, R., Basi, G.S., Caccavello, R.,
Davis, D., Doan, M., Dovey, H.F., Frigon, N., Hong, J., et al.
(1999). Purification and cloning of amyloid precursor protein b-
secretase from human brain. Nature 402, 537–540.
Taylor, A. (1993). Aminopeptidases: structure and function. FASEB
J. 7, 290–298.
Tedelind, S., Poliakova, K., Valeta, A., Hunegnaw, R., Yemanabe-
rhan, E.L., Heldin, N.E., Kurebayashi, J., Weber, E., Kopitar-
Jerala, N., Turk, B. et al. (2010). Nuclear cysteine cathepsin
variants in thyroid carcinoma cells. Biol. Chem. 391, 923–935.
Tomasselli, A.G., Qahwash, I., Emmons, T.L., Lu, Y., Leone, J.W.,
Lull, J.M., Fok, K.F., Bannow, C.A., Smith, C.W., Bienkowski,
M.J., et al. (2003). Employing a superior BACE1 cleavage
sequence to probe cellular APP processing. J. Neurochem. 84,
1006–1017.
Roberds, S.L., Anderson, J., Basi, G., Bienkowski, M.J., Branstetter,
D.G., Chen, K.S., Freedman, S.B., Frigon, N.L., Games, D., Hu,
K., et al. (2001). BACE knockout mice are healthy despite lack-
ing the primary b-secretase activity in brain: implications for
Alzheimer’s disease therapeutics. Hum. Mol. Genet. 10, 1317–
1324.
Sahasrabudhe, S.R., Brown, A.M., Hulmes, J.D., Jacobsen, J.S.,
Vitek, M.P., Blume, A.J., and Sonnenberg, J.L. (1993). Enzy-
matic generation of the amino terminus of the b-amyloid pep-
tide. J. Biol. Chem. 268, 16699–16705.
Saido, T.C. (1998). Alzheimer’s disease as proteolytic disorders:
anabolism and catabolism of b-amyloid. Neurobiol. Aging 19,
S69–S75.
Schechter, I. (1970). On the active sites of proteases. Cleavage of
peptide bonds involving D-alanine residues by carboxypeptidase
A. Eur. J. Biochem. 14, 516–520.
Turk, B., Turk, D., and Turk, V. (2000). Lysosomal cysteine prote-
ases: more than scavengers. Biochim. Biophys. Acta 1477,
98–111.
Schechter, I. (2005). Mapping the active site of proteases in the
1960s and rational design of inhibitors/drugs in the 1990s. Curr.
Protein Pept. Sci. 6, 501–512.
Schechter, I. and Berger, A. (1966a). Peptides of L- and D-alanine.
Synthesis and optical rotations. Biochemistry 5, 3362–3370.
Schechter, I. and Berger, A. (1966b). The hydrolysis of diastereo-
isomers of alanine peptides by carboxypeptidase A and leucine
aminopeptidase. Biochemistry 5, 3371–3375.
Schechter, I. and Berger, A. (1967). On the size of the active site
in proteases. I. Papain. Biochem. Biophys. Res. Commun. 27,
157–162.
Schechter, I. and Berger, A. (1968). On the active site of proteases.
3. Mapping the active site of papain; specific peptide inhibitors
of papain. Biochem. Biophys. Res. Commun. 32, 898–902.
Turner, R.T., Koelsch, G., Hong, L., Castanheira, P., Ermolieff, J.,
Ghosh, A.K., and Tang, J. (2001). Subsite specificity of mema-
psin 2 (b-secretase): implications for inhibitor design. Biochem-
istry 40, 10001–10006.
Vassar, R., Bennett, D., Babu-Khan, S., Kahn, S., Mendiaz, E.A.,
Denis, P., Teplow, D.B., Ross, S., Amarante, P., Loeloff, R., et
al. (1999). b-Secretase cleavage of Alzheimer’s amyloid precur-
sor protein by the transmembrane aspartic protease BACE. Sci-
ence 286, 735–741.
Wang, R., Sweeney, D., Gandy, S.E., and Sisodia, S. (1996). The
profile of soluble amyloid b protein in cultured cell media. J.
Biol. Chem. 271, 31894–31902.
Wolfe, M.S. (2006). Shutting down Alzheimer’s. Sci. Am. 294,
72–79.
Brought to you by | University of California - San Francisco
Authenticated
Download Date | 12/13/14 1:15 AM

Article in press - uncorrected proof
New b-secretase candidates rapidly cleave the WT b-site 569
Yan, R., Bienkowski, M.J., Shuck, M.E., Miao, H., Tory, M.C., Pau-
ley, A.M., Brashier, J.R., Stratman, N.C., Mathews, W.R., Buhl,
A.E., et al. (1999). Membrane-anchored aspartyl protease with
Alzheimer’s disease b-secretase activity. Nature 402, 533–537.
Yang, L.B., Lindholm, K., Yan, R., Citron, M., Xia, W., Yang, X.L.,
Beach, T., Sue, L., Won, P., Price, D., et al. (2003). Elevated b-
secretase expression and enzymatic activity detected in sporadic
Alzheimer disease. Nat. Med. 9, 3–4.
Received December 15, 2010; accepted February 5, 2011
Brought to you by | University of California - San Francisco
Authenticated
Download Date | 12/13/14 1:15 AM