Racemization of the aspartic acid residue of amyloid-β peptide by a radical reaction

Article abstract of DOI:10.1271/bbb.120797

Human amyloid-β peptide 1-42 (Aβ) was subjected to a radical reaction by using ascorbic acid and CuCl₂. The percentage of D-aspartic acid (D-Asp) after 24 h had increased to 6.69 ± 0.09%, this being comparable with the reported D-Asp concentration of purified core amyloids in Alzheimer's disease patients. This racemization was significantly inhibited by radical scavengers. L-Alanine was also racemized during the same reaction.

Full text of DOI:10.1271/bbb.120797

Biosci. Biotechnol. Biochem., 77 (2), 416–418, 2013
Note
Racemization of the Aspartic Acid Residue of Amyloid-ꢀ Peptide
by a Radical Reaction
1
1
1
1
Koharu TAMBO, Tomomi YAMAGUCHI, Keiko KOBAYASHI, Eri TERAUCHI,
2
1;3;y
Ikuyo ICHI, and Shosuke KOJO
1
Department of Food Science and Nutrition, Nara Women’s University, Nara 630-8506, Japan
Department of Food Science and Nutrition, Ochanomizu University, Bunkyo-ku, Tokyo 112-8610, Japan
The Open University of Japan, 2-11 Wakaba, Mihama-ku, Chiba 261-8586, Japan
2
3
Received October 15, 2012; Accepted October 30, 2012; Online Publication, February 7, 2013
[doi:10.1271/bbb.120797]
Human amyloid-ꢀ peptide 1-42 (Aꢀ) was subjected to
(ThT) to a final Aꢀ concentration of 20 mM. The radical
a radical reaction by using ascorbic acid and CuCl2. The
percentage of D-aspartic acid (D-Asp) after 24 h had
increased to 6:69 Æ 0:09%, this being comparable with
the reported D-Asp concentration of purified core
amyloids in Alzheimer’s disease patients. This racemi-
zation was significantly inhibited by radical scavengers.
L-Alanine was also racemized during the same reaction.
reaction was initiated by adding a CuCl solution (10 mM
2
final concentration) and ASC solution (1 mM final
concentration). This reaction was performed at room
temperature while agitating with a magnetic stirrer.
Reactions without copper ions [Cu (II)], without ASC,
and without both Cu (II) and ASC were also performed.
At 0, 1.5, 3, 6, 12, 24, and 48 h after initiating the
reaction, 150-mL aliquots were taken and mixed with
1.5 mL of an ethylenediaminetetraacetic acid (EDTA)
solution (5 mM, pH 7.4) to terminate the reaction. A
Key words: Alzheimer’s disease; amyloid; D-aspartic
acid; racemization; radical
100-mL aliquot was taken from this mixture, quickly
Senile plaques in the cerebral parenchyma are typical
features of Alzheimer’s disease (AD), and primarily
consist of aggregated fibrillar amyloid-ꢀ peptides.
Amyloid-ꢀ peptides have 39–42 residues, among which
amyloid-ꢀ peptide 1-42 (Aꢀ) is a major component
evaporated to dryness, and subjected to a determination
of D-amino acids. Another 35-mL aliquot was diluted 10-
fold with a 20 mM HEPES buffer (pH 7.4) containing
100 mM NaCl and 3.3 mM ThT, and the fluorescence
intensity of ThT was determined to evaluate fibril
1
)
9)
of cerebrovascular amyloid deposits. L-Aspartic acid
L-Asp) residues of Aꢀ are partially converted to D-Asp,
formation, using an RF-1500 fluorescence spectropho-
tometer (Shimadzu, Kyoto, Japan).
(
and this racemization of Aꢀ may hamper the proteolytic
attack on Aꢀ by aminopeptidase, resulting in the
The residue that had been prepared as already
described was subjected to hydrolysis under nitrogen
in a JLC007363 reaction vial (Waters Corp., Tokyo,
Japan) by adding 100 mL of 6 N HCl containing 1% of
phenol. The reaction mixture was quickly evaporated
after hydrolysis, and the residue was diluted with 0.1%
2
,3)
stabilization of Aꢀ deposits in stable plaques.
Furthermore, Aꢀ containing D-Asp has caused apoptosis
in PC12 cells,4 and produced a toxic peptide that
)
5
)
damaged neurons.
An ionic reaction via a succinimide intermediate has
1
0)
HCl. D-Amino acids were determined as described.
6
)
been proposed as a mechanism for racemization.
However, oxidative stress has been implicated in the
The mean value of duplicated determinations was
recorded, four independent reactions being performed.
Multiple comparisons were made by a one-way analysis
of variance by the Tukey-Kramer post hoc test, p < 0:05
being considered statistically significant.
7
)
pathogenesis of AD, and Aꢀ has caused radical
reactions in the presence of copper ions and ascorbic
acid (ASC).8 If the hydrogen atom bound to the ꢁ-
carbon of L-Asp is abstracted by a radical, an Asp
)
Aꢀ was subjected to reactions with ASC (1 mM) and
2
carbon radical that is in a planar sp hybridization is
CuCl (10 mM). The ASC concentration in the rat brain is
2
1
1)
generated. When this carbon radical abstracts a hydro-
gen atom from a neighboring molecule to regenerate
Asp, the hydrogen atom comes from either the upper or
lower side, resulting in racemization. We attempted to
demonstrate that a radical reaction caused racemization
of L-Asp in Aꢀ.
Human amyloid-ꢀ peptide 1-42 (Aꢀ), purchased from
Peptide Institute (Osaka, Japan), was dissolved in 0.02%
NH4OH at a final concentration of 80 mM. This solution
was diluted with a 20 mM HEPES buffer (pH 7.4)
containing 100 mM NaCl and 3.3 mM of thioflavin T
approximately 2 mM, and the copper concentration in
the normal dry mouse brain is approximately 94 mM (i.e.,
1
2)
6.4 ppm). The concentrations of ASC and copper ions
used in this study may therefore be appropriate as an
in vivo model. After reacting for 3 h, the percentage of
D-Asp was 3:46 ꢀ 0:17%, being significantly higher than
that at 0 h (2:09 ꢀ 0:06%; Fig. 1). This initial value was
due to racemization during acid hydrolysis in the
procedure for determining D-Asp. The percentage of
D-Asp increased to 6:69 ꢀ 0:09% after 24 h, this being
its plateau level (Fig. 1). This value is comparable with
y
To whom correspondence should be addressed. Tel/Fax: +81-43-298-4691; E-mail: kojo@ouj.ac.jp
Abbreviations: AD, Alzheimer’s disease; Aꢀ, human amyloid-ꢀ peptide 1-42; ASC, ascorbic acid; Asn, asparagine; Asp, aspartic acid

Radical Racemization of Aspartic Acid in Amyloid-ꢀ
417
Fig. 2. Change in ThT Fluorescence during the Reaction of Amyloid-
ꢀ Peptide.
Amyloid-ꢀ peptide 1-42 was reacted under the designated
Fig. 1. Change in D-Aspartic Acid Contents by Radical Reactions.
Amyloid-ꢀ peptide 1-42 was reacted under the designated con-
ditions [ Cu (II) þ, ASC þ; Cu (II) ꢁ, ASC þ;
Cu (II) þ,
conditions [ Cu (II) þ, ASC þ; Cu (II) ꢁ, ASC þ;
Cu (II)
ASC ꢁ;
Cu (II) ꢁ, ASC ꢁ]. The percentage of D-aspartic acid
þ, ASC ꢁ; Cu (II) ꢁ, ASC ꢁ]. ThT fluorescence was determined
as described in the Materials and Methods section. Each value
represents the mean ꢀ SE from four independent experiments.
Asterisks indicate a significant difference from the initial value
(p < 0:01; Tukey-Kramer post hoc test).
was determined as described in the Materials and Methods section.
Each value represents the mean ꢀ SE from four independent
experiments. Asterisks indicate a significant difference from the
initial value (p < 0:01; Tukey-Kramer post hoc test).
the reported D-Asp content which was 4.57%1 in
3)
0:08% in the respective presence of Trolox (2 mM) and
mercaptoethanol (10 mM). This inhibition of the race-
mization of L-Asp and L-Asn by radical scavengers
clearly demonstrated that a radical reaction caused
racemization of these amino acids. Racemization of the
Asp residue in elastin by a radical reaction has been
reported by Kuge et al.
The increase in D-Ala was also determined in the
reaction of Aꢀ with ASC and CuCl . The percentage of
2
D-Ala had significantly increased to 3:16 ꢀ 0:14% after
6 h when compared with its initial value (1:21 ꢀ 0:05%).
The percentage of D-Ala had increased to 3:57 ꢀ 0:14%
after 12 h, to 3:76 ꢀ 0:70% after 24 h, and then dropped
a little to 3:49 ꢀ 0:25% after 48 h. In the presence of
ASC alone, D-Ala had significantly increased to
1:39 ꢀ 0:08% after 24 h when compared with its initial
value (1:08 ꢀ 0:07%), and then increased further to
2:00 ꢀ 0:05% after 48 h, similar to the effect on Asp. In
the absence of ASC, D-Ala did not increase at all for
48 h from the starting point, either in the presence or
absence of Cu (II).
1
4)
purified core amyloids of AD patients and 7.31% in
the insoluble fraction of an AD patient’s brain.
It has been well established that Aꢀ, ASC, and Cu (II)
8
)
generated hydrogen peroxide and hydroxyl radicals,
making it possible that a radical reaction caused
racemization of Asp in Aꢀ.
1
5)
In the presence of ASC without Cu (II), the
percentage of D-Asp increased slightly, but significantly,
to 2:42 ꢀ 0:08% after 24 h and further increased to
3
:17 ꢀ 0:12% after 48 h (Fig. 1). ASC is a strong
reductant and easily reduces oxygen in the air to
generate superoxides which rapidly generate hydrogen
peroxide by dismutation, even in the absence of super-
oxide dismutase. The hydrogen peroxide thus formed
may have caused racemization, possibly aided by
contamination with trace transition metals in the glass
vessels used for this experiment. Although all the glass
vessels were washed with a nitric acid solution to
remove metals, these phenomena have often been
apparent in radical reactions. In the absence of ASC,
D-Asp did not increase at all for 48 h from the starting
point, either in the presence or absence of Cu (II)
D-Ala at 1.52% has also been detected in the insoluble
fraction of an AD patient’s brain, this being approx-
1
4)
(
Fig. 1). These results also show that racemization
imately one fifth of the percentage of D-Asp. The
proportion of D-Ala formed in this study was also much
less than that of D-Asp reflecting the higher reactivity
of Asp to radicals due to the electron-rich nature of its
ꢁ-carbon by the side-chain carboxylate anion when
compared with Ala.
caused by an ionic reaction for 48 h was negligible,
because ionic racemization could not appreciably pro-
ceed in the absence of ASC (an ionic reaction could of
course take place in the absence of ASC). The
racemization of Asp or Asn by a radical reaction is
much faster than that by an ionic reaction, even if it
took place.
Although inhibition experiments were not performed
because of the small change in D-Ala, a radical reaction
also seemed to be the mechanism for racemization,
Aꢀ was subjected to reactions with ASC (1 mM) and
6
)
CuCl (10 mM) for 24 h, after which the percentage of
because Ala cannot form a succinimide intermediate
during ionic racemization, this being different from Asp
and Asn.
2
D-Asp was 6:69 ꢀ 0:09%. The formation of D-Asp was
significantly inhibited to 5:49 ꢀ 0:10% and 2:95 ꢀ

4
18
K. TAMBO et al.
2) Roher AE, Lowenson JD, Clarke S, Wolkow C, Wang R, Cotter
The increase in ThT fluorescence was also determined
RJ, Reardon IM, Zuercher-Neely HA, Heinrikson RL, Ball MJ,
and Greenberg BD, J. Biol. Chem., 268, 3072–3083 (1993).
Iwatsubo T, Saido TC, Mann DMA, Lee VM-Y, and
Trojanowski JQ, Am. J. Pathol., 149, 1823–1830 (1996).
Kaneko I, Morimoto K, and Kubo T, Neurosci., 104, 1003–1011
(2001).
in the reaction of Aꢀ with ASC and CuCl2. Almost no
change was apparent in the ThT fluorescence. In the
absence of Cu (II) and ASC, the ThT fluorescence had
significantly increased as early as within 1.5 h, increased
further by 3 h, and reached a plateau after 6 h, before
gradually decreasing (Fig. 2). The change in ThT
fluorescence was similar in the presence of only ASC
3
4
)
)
5) Kubo T, Nishimura S, Kumagae Y, and Kaneko I, J. Neurosci.
Res., 70, 474–483 (2002).
6)
7)
8)
Geiger T and Clarke S, J. Biol. Chem., 262, 785–794 (1987).
Markesbery WR, Free Radic. Biol. Med., 23, 134–147 (1997).
Smith DG, Cappai R, and Barnham KJ, Biochim. Biophys. Acta,
(
Fig. 2). In the presence of only Cu (II), the ThT
fluorescence had significantly increased by 3 h and
continued to increase thereafter (Fig. 2). These results
clearly indicate that fibril formation was independent of
the racemization of L-Asp during the reaction for 48 h,
1
768, 1976–1990 (2007).
9
)
Naiki H, Higuchi K, Hosokawa M, and Takeda T, Anal.
Biochem., 177, 244–249 (1989).
consistent with the results of the study by Murakami
et al.16) Copper ions rather inhibited fibril formation, this
10) Aswad DA, Anal. Biochem., 137, 405–409 (1984).
11) Tokumaru S, Takeshita S, Nakata R, Tsukamoto I, and Kojo S,
J. Agric. Food Chem., 44, 2748–2753 (1996).
1
7)
being consistent with the results of Yoshiike et al.
Although aggregated fibrillar Aꢀ peptides are char-
acteristic of AD, the mean level of soluble Aꢀ has been
1
2) Rajendran R, Minqin R, Ynsa MD, Casadesus G, Smith MA,
Perry G, Halliwell B, and Watt F, Biochem. Biophys. Res.
Commun., 382, 91–95 (2009).
1
8)
highly correlated with markers of AD severity. In
addition, soluble Aꢀ, and not fibrillar Aꢀ, has caused
13) Shapira R, Austin GE, and Mirra SS, J. Neurochem., 50, 69–74
(1988).
1
9)
1
4) Mori H, Ishii K, Tomiyama T, Furiya Y, Sahara N, Asano S,
Endo N, Shiarsawa T, and Takio K, Tohoku J. Exp. Med., 174,
very early neuronal dysfunction. The role of soluble
Aꢀ, and particularly Aꢀ with D-amino acids, may
therefore be important for understanding the etiology
of AD.
2
51–262 (1994).
1
1
5) Kuge K, Kitamura K, Nakaoji K, Hamada K, Fujii N, Saito T,
and Fujii N, Chem. Biodivers., 7, 1408–1412 (2010).
6) Murakami K, Uno M, Masuda Y, Shimizu T, Shirasawa T,
and Irie K, Biochem. Biophys. Res. Commun., 366, 745–751
Acknowledgments
(
2008).
7) Yoshiike Y, Tanemura K, Murayama O, Akagi T, Murayama
1
1
1
This work was supported by grant-aid from the
Ministry of Education, Culture, Sports, Science and
Technology of Japan.
M, Sato S, Sun X, Tanaka N, and Takashima A, J. Biol. Chem.,
2
76, 32293–32299 (2001).
8) McLean CA, Cherny RA, Fraser FW, Fuller SJ, Smith MJ,
Beyreuther K, Bush AI, and Masters CL, Ann. Neurol., 46, 860–
866 (1999).
References
9) Busche MA, Chen X, Henning HA, Reichwald J, Staufenbiel M,
Sakmann B, and Konnerth A, Proc. Natl. Acad. Sci. USA, 109,
8740–8745 (2012).
1
)
Roher AE, Lowenson JD, Clarke S, Woods AS, Cotter RT,
Gowing E, and Ball MJ, Proc. Natl. Acad. Sci. USA, 90, 10836–
10840 (1993).