Structural studies of copper(I) complexes of amyloid-β peptide fragments: Formation of two-coordinate bis(histidine) complexes

Article abstract of DOI:10.1002/anie.200803908

(Chemical Equation Presented) The beta bind: Copper(I) binds to amyloid β-peptide fragments (see structure) as a stable bis(histidine), two-coordinate, near-linear complex, even in the presence of potential additional ligands. As has been proposed or assumed in other studies, the copper(I)-peptide complexes react with dioxygen to form the reactive oxygen species H₂O₂, without the need for a third histidine ligand to promote the chemistry.

Full text of DOI:10.1002/anie.200803908

Communications
DOI: 10.1002/anie.200803908
Peptide Chemistry
Structural Studies of Copper(I) Complexes of Amyloid-b Peptide
Fragments: Formation of Two-Coordinate Bis(histidine) Complexes**
Richard A. Himes, Ga Young Park, Gnana Sutha Siluvai, Ninian J. Blackburn, and
Kenneth D. Karlin*
Extensive evidence points to oxidative stress as a key event in
the pathogenesis and exacerbation of Alzheimerꢀs Disease
(AD).[1] Transition metals, such as Zn, Fe, and Cu, are present
in elevated concentrations in AD brain deposits, composed
primarily of 40- or 42-mer amyloid beta (Ab) peptides. The
redox-active copper(II) ion binds to the unstructured, hydro-
philic N terminus of Ab;[1g,2] and the ability of copper to
promote the formation of reactive oxygen species (ROS) and
cause neuronal death by interaction with Ab has been
demonstrated in vitro.[1a,c,3,4] ROS formation is proposed to
occur by interaction of reduced CuI–Ab with O2 or H2O2.
However, few direct studies of CuI binding or reactivity with
Ab peptides or fragments have been reported.[5,6]
We have studied the interactions of the hydrophilic N-
terminal region of the Ab peptide with CuI. An understanding
of the full redox competency of Cu–Ab, leading to ROS
formation and oxidative stress (that is, to cause events
associated with the onset of AD), is incomplete without
elucidation of the structure/function relationships of the
reduced (active) copper(I)–peptide complexes. We report
herein studies on the interaction of CuI ions with small
portions of the Ab peptide incorporating specific metal-
binding (His6, His13, His14) or potentially redox-active
(Tyr10) residues (Figure 1). Of considerable interest are the
contiguous His13 and His14 residues. We have previously
reported studies on CuI complexes of modified (by end-
capping and/or regiospecific Ne- or Nd-alkylation) His–His
Figure 1. Ab peptides used for studies with CuI ions.
A range of peptides (Figure 1) were synthesized and
purified by reverse-phase (RP) HPLC to a single peak. Their
identity and purity were confirmed by ESI mass spectrometry.
The peptides were stored either as lyophilized powders or as
stock solutions in doubly distilled deionized water, both at
ꢀ808C.[7] Copper(I)–peptide complexes were prepared
directly from CuI starting materials in the absence of
reductants, and their formulation confirmed using ESI-MS.
Structural information was obtained by spectroscopic techni-
ques for both solid and solution states (see below). Solid
samples of CuI–Ab(6–14) and CuI–Ab(10–14) were prepared
by incubating stoichiometric amounts of the respective
peptides with a [CuI(CH3CN)4]+ salt in DMF and isolated
by precipitation with diethyl ether, filtration, and drying
under reduced pressure. Their formulation was confirmed
using ESI-MS.[8] Mishandled samples turned deep blue,
indicating oxidation to CuII, whereas the CuI complexes
remained white-to-gray when air was excluded, indicating
reduced metal–peptide complexes.
dipeptides which, significantly, adopt a two-coordinate, near-
I
linear NHis Cu NHis environment.[6] In this report, we
demonstrate that CuI complexes of longer Ab peptide
fragments adopt the same apparent two-coordinate structure
in the solid state and aqueous solution. Preliminary reactivity
investigations, described here, indicate that the His13–CuI–
His14 moiety is the active part of the structure, responsible for
copper-Ab reactivity.
ꢀ
ꢀ
[*] Dr. R. A. Himes, G. Y. Park, Dr. K. D. Karlin
Dept. of Chemistry, The Johns Hopkins University
3400 N. Charles St., Baltimore, MD, 21218 (USA)
E-mail: kkarlin1@jhu.edu
G. S. Siluvai, Dr. N. J. Blackburn
For these solid samples, X-ray absorption spectroscopy
(XAS) was used as a powerful (yet unexploited, in the case of
Cu–Ab complexes) tool for the determination of oxidation
state, coordination environment, and bond lengths in the
derived metal complexes.[9,10] For both Ab(6–14) and Ab(10–
14) complexes, the occurrence of the 1s!4p transition at
8983–84 eV (Figure 2) definitively indicated that copper was
Dept. of Environmental and Biomolecular Systems
OGI School of Science and Engineering at OHSU
Beaverton, OR, 97006 (USA)
[**] This work was supported by the NIH (Grants GM28962, K.D.K.;
NIH Postdoctoral Fellowship, R.A.H.; NIH NS27583, N.J.B.).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200803908.
9084
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 9084 –9087

Angewandte
Chemie
FTIR spectroscopy. The stretching frequency of copper(I)-
bound CO is diagnostic for the overall coordination number
in cationic copper(I) species,[6,11e,12] and has been noted in
cuprous enzymes.[13] All three complexes had stretching
frequencies greater than 2110 cmꢀ1, varying by no more
than 2 cmꢀ1 (Table 1). The high frequency is clearly indicative
Table 1: Structural data for CuI complexes of His-containing peptides.
Donors[a]
Cu NImid [Š]
uCO[b] [cmꢀ1
]
ꢀ
Complex
[CuILd]+[c]
2 His
2 His
3 His
2 His
2 His
2 His
1 Imid
1.876
1.869
1.876
1.878
N/A
2110[d]
2105[e]
2110[f]
2112[f]
2110[f]
2075
[CuILH]+[c]
[CuIAb(6–14)]+
[CuIAb(10–14)]+
[CuIFHH]+
[CuILd(MeImid)]+[c]
1.896
2.008[g]
Figure 2. EXAFS (top, including insets) and XANES (bottom) spectro-
scopic data for CuI–Ab(6–14) (left) and CuI–Ab(10–14) (right). Fourier
transforms: black, fits: gray.
[a] N-Donor ligands available for coordination to CuI. [b] For correspond-
I
ꢀ
ing peptide–Cu CO complex. [c] Copper(I) complexes of His–His
dipeptides; Ld contains two trityl-protected imidazole
e nitrogen
atoms, whereas LH incorporates two unprotected imidazole moieties.
See Ref. [6]. [d] Dichloromethane solution. [e] Methanol solution.
in the + 1 oxidation state. Extended X-ray absorption fine
structure (EXAFS) spectroscopic data fits for the CuI–
Ab(10–14) complex, with only two histidine residues
(Figure 1), indicated two nitrogen ligands from imidazole
donors, as further supported by back-scattering from the ring
carbons and nitrogen. The data were consistent with these
donors being the only ligands bound to the CuI ion. The
intensity of the pre-edge (1s!4p) feature (Figure 2) was
further indicative of two-coordination, to the exclusion of
other (i.e., three-coordinate) geometries.[9,10] In addition, the
ꢀ
[f] With HEPES buffer, pH 7.4, D2O. [g] Two Cu NHis bonds 1.896 Š,
ꢀ
Cu NImid bond 2.008 Š.
of the presence of only two N donors coordinating to the CuI
ion. The results recalled our previous finding that His–His
dipeptide moieties strongly favor near-linear two-coordina-
tion (Table 1).[6]
The similarity in structure deduced for these complexes,
[CuI–Ab(6–14) and CuI–Ab(10–14), by EXAFS and IR
spectroscopy (CO binding); and also CuI(FHH), by IR]
strongly suggests that His13 and His14 constitute the two N-
donor ligands to the CuI center. Whereas the unique redox
properties of a CuI ion in a linear, two-coordinate environ-
ment have been noted in model complexes[11a,e,14] and the
structure has been proposed to be important in some copper-
enzyme active sites,[15] the possibility of a CuI(His)2 site
involved in Ab chemistry has been overlooked.
With our structural results in mind, we have begun
studying the redox reactivity of these systems. Preliminary
experiments on the ability of CuI–Ab fragment complexes to
produce ROS have been carried out. The first step in Cu–Ab
ROS production has been proposed to be CuII reduction
followed by reaction with O2 to produce H2O2.[16] Hydrogen
peroxide has been formed in vitro from Cu–Ab complexes,
but only in the presence of very large excesses of reducing
agents, such as ascorbate,[16,17] or by electrochemical reduction
of CuII.[4] Direct reactivity of CuI–Ab with O2, by way of the
reactions shown in Scheme 1, has not been studied, until now.
Production of H2O2 from oxygenated CuI–peptide solu-
tions was monitored using the horseradish peroxidase (HRP)/
Amplex Red assay. Hydrogen peroxide is produced from
solutions of CuI–Ab over the course of one hour, in amounts
significantly greater than CuI-only or peptide-only control
reactions.[8,18] Most intriguingly, all three systems, whether
incorporating the third His residue (His6) or not, or
incorporating the potentially redox-active Tyr10 or not,
produce assayable H2O2 in similar yields and rates of
ꢀ
short Cu N bond lengths—at 1.878 Š—are characteristic of
linear, two-coordinate geometry in copper(I)–nitrogen ligand
complexes, by comparison to crystallographically character-
ized synthetic copper(I) complexes.[11] The data also conform
to the structures identified previously in our CuI(His)2
dipeptide complexes, in which intramolecular binding of the
ꢀ
imidazole moieties of the dipeptide affords tight, linear Cu N
two-coordinate geometry.[6]
Further results obtained for CuI–Ab(6–14) firmly dem-
onstrate the propensity for CuI to adopt near-linear two-
coordinate geometry: EXAFS spectroscopic analysis of solid
CuI–Ab(6–14) indicated formation of the same structure,
despite the presence of a third potential histidine ligand. For
CuI–Ab(6–14), the Fourier Transform with fit is shown in
Figure 2. The data could only be fit to two N/O scatterers, thus
indicating the presence of only two ligands at the CuI center;
these were identified unambiguously as His nitrogen atoms by
ꢀ
backscattering. The Cu NHis bond lengths of 1.876 Š and the
X-ray absorption near-edge structure (XANES) absorption
intensity clearly indicate two-coordination (and three-coor-
dination).
Binding of CO to CuI was used as a probe of solution
structure. Results indicated that the 2Nimid structure persists in
solution, even for the three-His-containing complex CuI–
Ab(6–14). CO complexes were formed for each of the three
peptides [Figure 1: Ab(6–14) and Ab(10–14), discussed
above, and the tripeptide FHH, discussed in more detail
below]in 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
(HEPES) buffered (pH 7.4) D2O and characterized using
Angew. Chem. Int. Ed. 2008, 47, 9084 –9087
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 9085

Communications
behavior (H2O2 production) of CuIAb, or that a third His in
the sequence (His6) may not be crucial. All the current
literature suggests His6 as a ligand for the oxidized CuII ion.
These studies, absent of any CuI–Ab structural information,
conclude that a three-histidine binding environment is
probably important for the CuII/CuI-promoted production of
ROS. We have shown herein that this may not be the case. AD
oxidative stress chemistry is dependent upon: 1) the CuII/CuI
redox cycle, and 2) ROS production from CuI/H2O2 and/or
CuI/O2 chemistry. The energetics and kinetics of both may be
highly tuned by the preferred stable copper(I)–bis(histidine)
structure, which we have previously demonstrated has unique
redox properties.[6] As such, the study of CuI–Ab may hold
additional information, key to the understanding of Cu–Ab
oxidative stress.
Scheme 1. Potential reactions of Cu–Ab with dioxygen to form H2O2.
formation. Mechanistically, reduction of O2 to H2O2 requires
two electrons (Scheme 1). Thus, in the absence of an
exogenous reductant (as in these experiments), stoichiometry
requires that a second electron must be provided either by a
second copper ion in the CuI–Ab moiety or by the peptide
itself, potentially by tyrosine oxidation [Eq. (2)].
Based on our results (Figure 3), the similar efficiency of
CuI(FHH) in H2O2 production, compared to that of the Tyr-
containing species, suggests that electrons are supplied only
Experimental Section
Experimental procedures, including procedures for peptide synthesis
and purification, preparation of solid and solution CuI–peptide
samples, procedures for CO-binding and H2O2-producing experi-
ments, and methods for EXAFS spectroscopic data collection and
analysis, are available in the Supporting Information.
Received: August 7, 2008
Published online: October 17, 2008
Keywords: amyloids · copper · EXAFS spectroscopy ·
.
hydrogen peroxide · reactive species
Figure 3. Yields of H2O2 from reactions of O2 with 25 mm copper(I)–
peptide solutions, as determined by HRP/Amplex Red assay. CuI
complex: black, peptide-only: gray. A) Ab(6–14); B) Ab(10–14);
C) FHH. Error bars represent standard errors from five trials.
[1]a) K. J. Barnham, A. I. Bush, Curr. Opin. Chem. Biol. 2008, 12,
222 – 228; b) R. R. Crichton, D. T. Dexter, R. J. Ward, Coord.
Chem. Rev. 2008, 252, 1189 – 1199; c) P. J. Crouch, S. M. E.
Harding, A. R. White, J. Camakaris, A. I. Bush, C. L. Masters,
Int. J. Biochem. Cell Biol. 2008, 40, 181 – 198; d) A. Rauk, Dalton
Trans. 2008, 1273 – 1282; e) L. M. Sayre, G. Perry, M. A. Smith,
Chem. Res. Toxicol. 2008, 21, 172 – 188; f) A. Nunomura, R. J.
Castellani, X. W. Zhu, P. I. Moreira, G. Perry, M. A. Smith, J.
Neuropathol. Exp. Neurol. 2006, 65, 631 – 641; g) E. Gaggelli, H.
Kozlowski, D. Valensin, G. Valensin, Chem. Rev. 2006, 106,
1995 – 2044.
[2]a) V. A. Streltsov, S. J. J. Titmuss, V. C. Epa, K. J. Barnham, C. L.
Masters, J. N. Varghese, Biophys. J. 2008, 95, 3447 – 3456; b) J. W.
Karr, V. A. Szalai, Biochemistry 2008, 47, 5006 – 5016; c) C. D.
Syme, R. C. Nadal, S. E. J. Rigby, J. H. Viles, J. Biol. Chem. 2004,
279, 18169 – 18177.
by the oxidation of copper. Furthermore, the similar rates and
yields among all three species suggest His6 is not significantly
involved in CuI–Ab–O2 reactivity. In other words, the
uniformity in results from these preliminary experiments
with the three copper–peptide species suggests that they react
with O2 to produce H2O2 by the same mechanism—Equa-
tion (1), wherein two separate Ab–CuI moieties are involved
and each CuI–Ab species is a CuI(His)2 complex.[19] Together,
these results suggest that the CuI(His)2 unit may not only be
the predominant binding mode of CuI ions to Ab peptides, but
also the structure directly responsible for the behavior
(including ROS production) of reduced CuI–Ab species.
In summary, our EXAFS spectroscopy and CO-binding
studies have clearly demonstrated the preference of CuI ions
for two-coordinate geometry in binding to fragments of the
Ab N-terminal region through a contiguous His13–His14
motif. That this structure is retained, even in the presence of
three histidine residues (His6, His13, His14) and additional
potential donors (Tyr10, Asp7, Glu11, Ser8, backbone
carbonyl O, amide N), is striking. The two-coordinate geom-
etry of CuI–Ab may prove critical to understanding the redox
chemistry of Cu–Ab, and thus to understanding oxidative
stress in AD. Preliminary ROS results indicate that the two-
coordinate CuI(His)2 structure is significant for explaining the
[3]P. S. Donnelly, Z. Xiao, A. G. Wedd, Curr. Opin. Chem. Biol.
2007, 11, 128 – 133.
[4]D. L. Jiang, L. J. Men, J. X. Wang, Y. Zhang, S. Chickenyen, Y. S.
Wang, F. M. Zhou, Biochemistry 2007, 46, 9270 – 9282.
[5]a) R. Baruch-Suchodolsky, B. Fischer, Biochemistry 2008, 47,
7796 – 7806; b) V. A. Streltsov, J. N. Varghese, Chem. Commun.
2008, 3169 – 3171; c) D. F. Raffa, G. A. Rickard, A. Rauk, J. Biol.
Inorg. Chem. 2007, 12, 147 – 164.
[6]R. A. Himes, G. Y. Park, A. N. Barry, N. J. Blackburn, K. D.
Karlin, J. Am. Chem. Soc. 2007, 129, 5352 – 5353.
[7]Periodic reverse-phase HPLC confirmed that no peptide
decomposition occurred when stored in this manner for weeks.
No aggregation or precipitation occurred.
[8]See Supporting Information.
[9]N. J. Blackburn, R. W. Strange, J. Reedijk, A. Volbeda, A.
Farooq, K. D. Karlin, J. Zubieta, Inorg. Chem. 1989, 28, 1349 –
1357.
9086 www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 9084 –9087

Angewandte
Chemie
[10]L. S. Kau, D. J. Spira-Solomon, J. E. Pennerhahn, K. O. Hodg-
son, E. I. Solomon, J. Am. Chem. Soc. 1987, 109, 6433 – 6442.
[11]a) I. Sanyal, K. D. Karlin, R. W. Strange, N. J. Blackburn, J. Am.
Chem. Soc. 1993, 115, 11259 – 11270; b) A. Habiyakare, E. A. C.
Lucken, G. Bernardinelli, J. Chem. Soc. Dalton Trans. 1991,
2269 – 2273; c) M. Munakata, S. Kitagawa, H. Shimono, H.
Masuda, Inorg. Chim. Acta 1989, 158, 217 – 220; d) L. M.
Engelhardt, C. Pakawatchai, A. H. White, P. C. Healy, J. Chem.
Soc. Dalton Trans. 1985, 117 – 123; e) T. N. Sorrell, D. L.
Jameson, J. Am. Chem. Soc. 1983, 105, 6013 – 6018; f) Y.
Agnus, R. Louis, R. Weiss, J. Chem. Soc. Chem. Commun.
1980, 867 – 869; g) A. H. Lewin, I. A. Cohen, R. J. Michl, J.
Inorg. Nucl. Chem. 1974, 36, 1951 – 1957; h) H. Okkersen,
Groeneve. Wl, J. Reedijk, Recl. Trav. Chim. Pays-Bas 1973, 92,
945 – 953.
[12]a) A. P. Cole, V. Mahadevan, L. M. Mirica, X. Ottenwaelder,
T. D. P. Stack, Inorg. Chem. 2005, 44, 7345 – 7364; b) C. C. Chou,
C. C. Su, A. Yeh, Inorg. Chem. 2005, 44, 6122 – 6128; c) J. K. Voo,
K. C. Lam, A. L. Rheingold, C. G. Riordan, J. Chem. Soc. Dalton
Trans. 2001, 1803 – 1805; d) Y. Rondelez, O. SØneque, M.-N.
Rager, A. F. Duprat, O. Reinaud, Chem. Eur. J. 2000, 6, 4218 –
4226; e) L. Casella, M. Gullotti, G. Pallanza, L. Rigoni, J. Am.
Chem. Soc. 1988, 110, 4221 – 4227; f) M. Pasquali, C. Floriani, A.
Gaetanimanfredotti, Inorg. Chem. 1980, 19, 1191 – 1197.
[13]S. Jaron, N. J. Blackburn, Biochemistry 1999, 38, 15086 – 15096.
[14]L. Le Clainche, M. Giorgi, O. Reinaud, Eur. J. Inorg. Chem.
2000, 1931 – 1933.
b) X. D. Huang, C. S. Atwood, M. A. Hartshorn, G. Multhaup,
L. E. Goldstein, R. C. Scarpa, M. P. Cuajungco, D. N. Gray, J.
Lim, R. D. Moir, R. E. Tanzi, A. I. Bush, Biochemistry 1999, 38,
7609 – 7616.
[17]X. D. Huang, et al., J. Biol. Chem. 1999, 274, 37111 – 37116.
[18]a) When carried out by the same method as the copper(I)–
peptide experiments[8] copper(I)-only control reactions give a
yield of H2O2 that never exceeds (and is often less than) 40% of
that detected for copper(I)–peptide (that is, < 1.5 mm, compared
to > 4 mm for copper(I)–peptide). It should be noted that there is
some unexplained dependence on experimental conditions,
especially the order of addition of reagents; for example, if CuI
solutions are added to aerobic, buffered HRP/Amplex Red
assay, strong signalling of H2O2 is indicated; b) if CuI ions are in
excess (2:1) relative to the peptide, the results are unchanged,
qualitatively indicating that CuI ion binds the peptide strongly.
Quantitative determinations are in progress; c) assayed yields of
H2O2 are, at most, approximately 30%, considering background
and small amounts of Amplex oxidation by copper(II)–peptide
and copper(I)–peptide/O2 (data not shown). It is unknown
whether relative inefficiency of HRP/Amplex Red trapping is
responsible for the sub-stoichiometric yield, or if the copper(I)/
(II)–peptide itself consumes some peroxide. Catalase attenuates
the signal. We are currently carrying out experiments to “trace”
all copper(I) electron equivalents.
[19]A recent publication implicates a soluble “dimer” (with two A b
moities and, thus, possibly two Cu ions) as the minimal unit
responsible for AD toxicity. See G. M. Shankar, S. Li, T. H.
Mehta, A. Garcꢁa-Muꢂoz, N. E Shepardson, I. Smith, F. M.
Brett, M. A. Farrell, M. J. Rowan, C. A. Lemere, C. M. Regan,
D. M. Walsh, B. L. Sabatini, D. J. Selkoe, Nature Medicine 2008,
14, 837 – 842.
[15]N. J. Blackburn, F. C. Rhames, M. Ralle, S. Jaron, J. Biol. Inorg.
Chem. 2000, 5, 341 – 353.
[16]a) C. Opazo, X. D. Huang, R. A. Cherny, R. D. Moir, A. E.
Roher, A. R. White, R. Cappai, C. L. Masters, R. E. Tanzi, N. C.
Inestrosa, A. I. Bush, J. Biol. Chem. 2002, 277, 40302 – 40308;
Angew. Chem. Int. Ed. 2008, 47, 9084 –9087
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 9087