Antibody recognition and conformational flexibility of a plaque-specific β-amyloid epitope modulated by non-native peptide flanking regions

Article abstract of DOI:10.1021/jm070196e

Here we report on the synthesis, antibody binding, and QSAR studies of a series of linear and cyclic peptides containing a β-amyloid plaque-specific epitope (Aβ(4-10); FRHDSGY). In these constructs, two or three α-L-Ala, α-D-Ala, or β-Ala residues were introduced at both N- and C-termini of the epitope as non-native flanking sequences. Cyclization of the linear Aβ(4-10) epitope peptide resulted in reduced antibody binding. However, the antibody binding could be fully compensated by insertion of alanine flanks into the corresponding cyclic peptides. These results indicate that the modification of a β-amyloid plaque-specific epitope by combination of cyclization and flanking sequences could generate highly antigenic peptides compared to the native sequence. A novel 3D QSAR method, which explicitly handles conformational flexibility, was developed for the case of such molecular libraries. This method led to the prediction of the binding conformation for the common FRHDSGY sequence.

Full text of DOI:10.1021/jm070196e

1150
J. Med. Chem. 2008, 51, 1150–1161
Antibody Recognition and Conformational Flexibility of a Plaque-Specific ꢀ-Amyloid Epitope
Modulated by Non-native Peptide Flanking Regions
Marilena Manea,†,‡ Adrián Kalászi,‡,§,4 Gábor Mezo˝, Kata Horváti, Andrea Bodor,# Anikó Horváth, Ödön Farkas,§,4
András Perczel,4, Michael Przybylski,† and Ferenc Hudecz*,4,
Laboratory of Analytical Chemistry and Biopolymer Structure Analysis, UniVersity of Konstanz, 78457 Konstanz, Germany, Laboratory of
Chemical Informatics, Institute of Chemistry, EötVös L. UniVersity, 1117 Budapest, Hungary, Research Group of Peptide Chemistry, Hungarian
Academy of Sciences, EötVös L. UniVersity, 1117 Budapest, Hungary, Department of Organic Chemistry, EötVös L.
UniVersity, 1117 Budapest, Hungary, Laboratory of Structural Chemistry and Biology, Institute of Chemistry, EötVös L. UniVersity,
1117 Budapest, Hungary, and Protein Modelling Group, Hungarian Academy of Sciences, Institute of Chemistry, EötVös L. UniVersity,
1117 Budapest, Hungary
ReceiVed February 20, 2007
Here we report on the synthesis, antibody binding, and QSAR studies of a series of linear and cyclic peptides
containing a ꢀ-amyloid plaque-specific epitope (Aꢀ(4–10); FRHDSGY). In these constructs, two or three
R-L-Ala, R-D-Ala, or ꢀ-Ala residues were introduced at both N- and C-termini of the epitope as non-native
flanking sequences. Cyclization of the linear Aꢀ(4–10) epitope peptide resulted in reduced antibody binding.
However, the antibody binding could be fully compensated by insertion of alanine flanks into the
corresponding cyclic peptides. These results indicate that the modification of a ꢀ-amyloid plaque-specific
epitope by combination of cyclization and flanking sequences could generate highly antigenic peptides
compared to the native sequence. A novel 3D QSAR method, which explicitly handles conformational
flexibility, was developed for the case of such molecular libraries. This method led to the prediction of the
binding conformation for the common FRHDSGY sequence.
Introduction
synthesis of cyclic and/or chimeric peptides;8–11 (ii) modification
of the flanking regions connected to the N- and/or C-terminus
of a core epitope;12,13 and (iii) multiplication of copies of a
defined number of B- or T-cell epitopes by conjugation to carrier
molecules.14 Recent work in our laboratories has been focused
on the preparation and characterization of various bioconjugates
as potential lead structures for AD vaccination.15 Thus, the
Aꢀ(4–10) epitope was elongated by either an N-terminal -Cys-
(Gly)5 - or a C-terminal -(Gly)5-Cys- sequence, and it was
attached via thioether linkage to different polylysine-based
branched chain polypeptide carriers with either Ser (SAK) or
Glu (EAK) residues at the end of the branches.15 In a second
type of conjugates, the Aꢀ(4–10) epitope alone, or flanked by
R-L- or ꢀ-alanine dimers at the N- and C-terminal sides, was
coupled via amide bonds to (i), a tetratuftsin derivative (Ac-
[TKPKG]4-NH2), or (ii), to a carrier peptide elongated by a
promiscuous helper T-cell epitope (Ac-FFLLTRILTIPQSLD-
[TKPKG]4-NH2).16 The evaluation of these conjugates showed
that the type of the carrier, the epitope topology, the presence
of a spacer group between the epitope and carrier, and of
flanking regions all had significant effects on the antibody
recognition of the Aꢀ(4–10) epitope. In particular, the introduc-
tion of ꢀ-alanine residues as flanks to both N- and C-termini of
the epitope provided a marked increase of antibody binding.
One of our previous studies demonstrated that the replacement
of amino acids by their D isomers at the N-terminus of an epitope
peptide increased its enzymatic stability in human sera.12 In
contrast, the substitution of the L-amino acids by the D form at
the C-terminus prevents the enzyme digestion in lysosomal
fractions. We could conclude that the incorporation of two to
three D-amino acid residues to both N- and C-termini can result
in high stability of epitope peptides in complex biological
matrices. In another experiment, we found that the cyclization
via disulfide bridge formation between cysteine residues attached
to the N- and C-terminal positions may increase the stability of
The accumulation and fibrillar association in plaques of
ꢀ-amyloid (Aꢀ), a self-aggregating peptide of 39–43 residues,
in brain is generally thought to be the cause of cognitive decline
and neurodegeneration in Alzheimer’s disease (AD)a.1–3 Re-
cently, possible therapeutic approaches for treatment of AD have
been pursued, aimed at the enhancement of amyloid clearance
via active or passive immunization by anti-Αꢀ antibodies. In
transgenic mouse models of AD, several laboratories have
demonstrated that therapeutically effective antibodies raised
against Aꢀ(1–42) and/or its oligomeric assemblies decrease Aꢀ
fibrillogenesis and cytotoxicity and are capable to exert thera-
peutic effects.1–6 We found previously that these therapeutically
active antibodies specifically target the N-terminal Aꢀ(4–10)
epitope sequence FRHDSGY.7 The recognition specificity and
properties of this epitope and derived peptides may provide a
lead structure both for development of new specific AD vaccines
and for molecular diagnostic applications.
Several approaches have been developed in order to modify
the immunorecognition of linear peptides representing B- or
T-cell epitopes and to increase their enzymatic stability: (i)
* To whom correspondence should be addressed at the Hungarian
Academy of Sciences. Phone: (+36)-1-209-0555. Fax: (+36)-1-372-2620.
E-mail: fhudecz@ludens.elte.hu.
†
Laboratory of Analytical Chemistry and Biopolymer Structure Analysis,
University of Konstanz.
‡
M.M. and A.K. contributed equally to this work.
Laboratory of Chemical Informatics, Eötvös L. University.
Department of Organic Chemistry, Eötvös L. University.
§
4
Research Group of Peptide Chemistry, Eötvös L. University.
Laboratory of Structural Chemistry and Biology, Eötvös L. University.
#
Protein Modelling Group, Eötvös L. University.
a Abbreviations: AD, Alzheimer’s disease; Aꢀ, ꢀ-amyloid peptide; a,
R-D-alanine; QSAR, quantitative structure-activity relationship; BC, binding
conformation; RCs, reference conformers, ACs, anchor conformations; MD,
molecular dynamics.
10.1021/jm070196e CCC: $40.75
2008 American Chemical Society
Published on Web 02/20/2008

Conformational Flexibility of a Plaque-Specific ꢀ-Amyloid Epitope
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5 1151
the biological active conformation of the epitope sequence.
However, the enzymatic stability of the peptide increased only
moderately in human sera compared to the linear epitope.11
These results prompted us to evaluate systematically the effects
of R-L-, ꢀ-, and R-D-alanine flanking regions on antigenicity
and conformation of linear and cyclic peptides containing the
ꢀ-amyloid(4–10) epitope.
In the present study, we found that the presence of non-native
L- and D-alanine (the latter will be marked in the text with “a”
as one letter code) flanking sequences caused a significant
increase of antibody recognition in linear ꢀ-amyloid(4–10)
epitope peptides, while cyclization of the native Aꢀ(4–10)
epitope resulted in lower antibody binding. However, this
decrease of antibody recognition in cyclopeptides could be
compensated by flanking alanine dimers or trimers at both N-
and C-termini. We also show that the type of flanking regions
has a significant effect on the conformation of both linear and
cyclic peptides, the presence of L-Ala2 and L-Ala3 or ꢀ-Ala2
and ꢀ-Ala3 sequences lead to increased conformational flexibility
in the core epitope region.
To gain insight into the conformational properties of the
analyzed peptides and to predict the binding conformation, a
novel 3D QSAR computation was performed. 3D QSAR
methods that account for the conformational flexibility have been
evolved17–22 during the past decade; however, the structures
examined in most cases contain few rotable bonds, and the
conformational space is usually sampled by some representative
low energy conformers. Usually the bioactive conformation is
a priori unknown, which may result in poor predictive capabil-
ity. On the other hand, there is a limited number of efforts to
develop methods specially designed to predict the binding
conformation for molecules containing substantial number of
rotable bonds.23,24 Such generally applicable computational tool
was developed and its operation is demonstrated on the peptides
analyzed in this study.
concentration of BC in solution, [IBC], depends only on ∆Gconf
(eqs 2 and 3).
[IR] ) [I0]e-ꢀ(∆G
[IBC] ) [I0]e-ꢀ(∆G
(2)
(3)
conf+∆Gbind
)
)
conf
where [I0] is the concentration of the most stable conformer in
solution. The fraction of [IR] and [IBC] depends only on the
∆Gbind value (eq 4). If the quotient is expanded by the total
concentration of the inhibitor (in our case the epitope peptide)
in solution, Itot, the relative concentration or molfraction of the
bound conformer is obtained in the numerator. We may also
replace the relative concentration of the binding conformation
in solution, the denominator, by the proportion of the conforma-
tions in a simulated trajectory, as the quotient of the frequency
of BC (fBC) and the total number of steps, ntraj, in a trajectory
which adequately represents the conformational distribution of
the flexible molecule in solution.
[IR]
[IR]
[IBC
]
]
Itot
Itot
) e-ꢀ∆G
)
)
(4)
bind
[IBC
]
fBC
[IBC
Itot
ntraj
The free energy change associated to the rigid binding can
be considered as the sum of the energy changes due to the
conformational change of the active site, the desolvation effect,
the formation of the receptor–ligand interaction, and the freezing
of the translational and rotational degrees of freedom of the
ligand during binding. These members can be regarded as
constant for the peptides of this set as these molecules are
specially designed to bind only via the CORE region. Hence,
the ∆Gbind values for each molecule are considered to be equal.
∆Gibind ) ∆Gbjind
(5)
It follows from eqs 4 and 5 that the ratios of the relative
concentrations of bound molecules for any member pairs of the
molecular library should be close to the ratios of the probabilities
of their BCs observed via the MD simulations in solution.
Theoretical Background
The analyzed molecules have a significant number of rotable
bonds and have a common structural feature, the FRHDSGY
epitope motif, which interacts with the antibody-binding site
(paratope). We designate this active core region as CORE. The
off-CORE parts in each single molecule vary from each other,
and the working hypothesis is that their direct interaction with
the binding site can be neglected by definition. The off-CORE
exhibits an influence on the binding activity indirectly, only by
modifying the conformational behavior of the flexible CORE.
The major goal of our novel 3D QSAR method is to locate the
binding conformation of the CORE, which can be found in the
bound state. In this process, no information about the 3D
structure of the active site is known, but the experimental activity
of each molecule is used.
Consider the process of binding of a flexible molecule to the
active site of a target protein. The binding free energy, ∆G bind-
tot, is independent of the actual reaction path, so we may split it
into two parts.25 The first part can be the formation of the binding
conformation (BC) in solution and the corresponding free energy
change is denoted as ∆Gconf. The second part is the “rigid” binding
of the BC to the active site, the corresponding free energy change
f i
ni
[IRi] [IRj]
[IRi]
Itot
[IRj]
Itot
Itot
f iBC
ni
Itot
)
)
(6)
j
j
f
f
nj
nj
The relative concentration of the bound molecules can be
determined from the results of well established experimental
methods. In the present case, we used the I50 values obtained
from an indirect ELISA experiment (I50 represents the antigen
concentration needed for obtaining an OD ) 1 at 450 nm).
During this experiment, the binding of the ligand, eq 7,
competes with the binding of the inhibitor or epitope peptide,
eq 8.
[L][R]
Kd )
Ki )
[L] ≈ Ltot
[I] ≈ Itot
(7)
(8)
[LR]
[I][R]
[IR]
is, ∆Gbind
.
The total concentration of the active site, Bmax, is expressed
in eq 9:
∆Gbind-tot ) ∆Gconf + ∆Gbind
(1)
[L] [I]
The concentration of the complex, [IR], depends on the free
energy change related to the whole binding process, while the
Bmax ) [R] + [IR] + [LR] ) 1 +
+
[R] (9)
(
)
Kd
Ki

1152 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5
Manea et al.
The relationship between the relative concentration of the
bound conformer and the I50 values, eq 11 can be obtained using
the Cheng-Prusoff equation,26 eq 10:
the nature of the bonds introduced for cyclization might also
affect the antibody recognition, and a fairly rigid amide bond
between N- and C-terminal may abolish the access to a binding
conformation. On the other hand, cyclic peptides containing
disulfide bridge have been shown to yield higher antigenicity
than the linear peptides.38
Based on these previous findings, the influence of R-L-, ꢀ-
and R-D-alanine flanking regions on antibody binding and
conformational flexibility was studied by the application of linear
and cyclic peptide analogs of the ꢀ-amyloid(4–10) epitope
sequence. Cyclization via disulfide bridge formation was
performed by dissolving the peptides in 50 mM aqueous
ammonium acetate, pH 6: dimethyl sulfoxide, 1:1 (v/v), and
the reaction was carried out at 25 °C for 24 h. In order to avoid
oligomerization of linear precursors, reactions were generally
performed at low peptide concentrations (0.2 mg/mL). The
cyclic products (as well as linear peptides) were characterized
by analytical HPLC and mass spectrometry, which ascertained
high purities and correct structures for all peptide derivatives
(Table 1 and Figure 1).
[L]
I50 ) KI 1 +
(10)
(
)
Kd
Bmax
I50[R]
Bmax
[IR] [IR]
Θ
)
)
)
)
)
Itot
I0
I50KI
[L]
Kd
2I50 I50
1 +
KI + I50
(
)
(11)
where
Bmax
Θ )
2
Θ values, as defined in eq 11, can be kept constant during
the experimental study; thus, they can be canceled out during
the correlation between the library members (eq 6). The term
1/I50 is referred in the following as binding activity. Combining
eqs 6 and 11 leads to eq 12.
The Effect of Flanking Sequences on Antibody Binding
of Linear and Cyclic Peptides. Linear and cyclic peptides
comprising the ꢀ-amyloid(4–10) epitope were compared for
binding to an anti-Aꢀ(1–17) monoclonal antibody by indirect
enzyme-linked immunosorbent assay (ELISA). In order to
overcome the drawbacks associated with the direct adsorption
of small peptides to coating surfaces due to differential coating
properties and possible conformational changes of the epitope,39
the peptides were attached to the surface of the ELISA plate
using the biotin-streptavidin interaction. To this purpose, a penta-
glycine spacer was introduced at the N-terminus of the peptides
in order to preserve the peptide conformation and improve the
epitope accessibility. Based on the ELISA data (Table 1, Figures
2 and 3), the effects of cyclization and/or flanking regions
adjacent to the B-cell epitope on the antibody binding could be
reproducibly assessed.
The cyclization of Aꢀ(4–10) epitope (16) resulted in a reduced
antibody binding compared to the linear peptide (1), while the
reverse sequence did not show any affinity to the monoclonal
antibody (mAb) (Table 1). Interestingly, the addition of two
and three R-L-alanine residues as flanking sequences to the
epitope led to substantially increased binding to the antibody,
both with the linear and cyclic antigens. The highest binding
was obtained for the linear peptide 3 in which the epitope was
flanked by two R-L-alanine residues, as compared to peptide 6
(epitope flanked by R-L-alanine trimers) and peptides 4 and 5
which contained ꢀ- or R-D-alanine flanking dimers (Figure 2
and Table 1).
Binding data for the cyclic peptides to the mAb are compared
in Figure 3. All peptides containing the Aꢀ(4–10) epitope with
alanine dimer or trimer showed high binding to the mAb,
compared to the cyclic peptide 16 without flanks. No significant
differences in the antibody binding were determined for the
cyclic peptides containing two or three R-L-alanine or ꢀ-alanine
residues. However, these peptides showed higher antigenicity
than the cyclic peptides 19 and 22 in which the epitope was
flanked by R-D-alanine dimers or trimers.
Effect of Flanking Sequences on the Conformation of
Linear and Cyclic Epitope Peptides Determined by Circular
Dichroism (CD) Spectroscopy. The conformational preferences
of linear and cyclic peptides containing the Aꢀ(4–10) epitope
flanked by R-L-, ꢀ-, or R-D-alanine dimers or trimers at both
N- and C-terminus were studied by CD spectroscopy. Spectra
were recorded in water and in 100% TFE, since the latter solvent
f i
[I5j0]
[Ii50]
fBC
[I50] ntraj
ni
f j
nj
1
)
(12)
Generally, we can conclude that for every member of a
suitable molecular library, the probability of the binding
conformation of the CORE in the unbound environment is a
linear function, with positive slope, of the binding activity. On
the other hand it has to be emphasized that the bound
conformation rarely coincides the lowest energy conformer in
the unbound state in solution. The number of rotable bonds of
the ligand primarily influences the strain energy as it was
demonstrated on several large scale studies27,28 on pharmaceuti-
cally relevant protein–ligand crystal complexes. The term strain
energy was used to the energy difference between the lowest
energy conformer in solution and the energy of the conformation
in the bound state. In these studies, not surprisingly, correlation
between bioactivity and the conformational flexibility was not
perceived while the ∆Gbind values varied for the molecules.
According to the general observation the flexible ligands tend
to bind in extended conformation, however a number of
exceptions as well as “maximally compact arrangements” was
also reported.29
In summary, the bioactive conformation does not necessary
prevail in the unbound state and “only” eq 12 should stand in
a specially designed library. A computational procedure was
developed, based on this equation, to locate the binding
conformation in the conformational ensemble in solution.
Results and Discussion
Synthesis and Characterization of Aꢀ-Epitope Peptides
Containing Flanking Sequences. The application of short linear
peptides as antigens has some well-known limitations, a major
drawback being the rapid proteolytic degradation in ViVo. The
incorporation of nonproteinogenic amino acids such as ꢀ-Ala30
or of D-amino acids31,32 is capable of providing protection from
proteolytic digestion. Another frequent problem of linear
peptides, namely their flexibility which may limit the presenta-
tion of binding conformations, has been shown to be overcome
by restricting peptide mobility through cyclization.33–36 How-
ever, cyclic peptides with small ring size might have low
antigenicity as compared to their linear versions.37 Furthermore,

Conformational Flexibility of a Plaque-Specific ꢀ-Amyloid Epitope
Table 1. Characteristics of Aꢀ(4–10) Epitope Peptides
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5 1153
peptide no.
sequencea
HPLC Rt (min)b–d
[M] calcd/founde,f
c (×10-8 mol/L)g
1.10 ( 0.12
1
2
3
4
5
6
7
8
biotin-GGGGGFRHDSGY-NH2
biotin-GGGGGYGSDHRF-NH2
24.80
25.16
25.26
24.52
25.32
25.42
24.17
25.41
27.34
27.16
25.47
27.08
28.06
25.11
25.51
25.37
25.53
24.07
25.37
25.13
23.66
25.11
21.19
21.83
22.37
23.39
1390.5837/1390.5755
1390.5837/1390.5792
1674.7321/1674.7292
1674.7321/1674.7305
1674.7321/1674.7285
1816.8063/1816.8089
1816.8063/1816.8133
1816.8063/1816.8001
1596.6020/1596.6168
1880.7505/1880.7446
1880.7505/1880.7459
1880.7505/1880.7455
2022.8247/2022.8350
2022.8247/2022.8197
2022.8247/2022.8218
1594.5863/1594.5868
1878.7348/1878.7340
1878.7348/1878.7293
1878.7348/1878.7368
2020.8090/2020.8117
2020.8090/2020.8160
2020.8090/2020.8023
921.4/921.8
biotin-GGGGGAAFRHDSGYAA-NH2
biotin-GGGGGꢀAꢀAFRHDSGYꢀAꢀA-NH2
biotin-GGGGGaaFRHDSGYaa-NH2
biotin-GGGGGAAAFRHDSGYAAA-NH2
biotin-GGGGGꢀAꢀAꢀAFRHDSGYꢀAꢀAꢀA-NH2
biotin-GGGGGaaaFRHDSGYaaa-NH2
biotin-GGGGGCFRHDSGYC-NH2
biotin-GGGGGCAAFRHDSGYAAC-NH2
biotin-GGGGGCꢀAꢀAFRHDSGYꢀAꢀAC-NH2
biotin-GGGGGCaaFRHDSGYaaC-NH2
biotin-GGGGGCAAAFRHDSGYAAAC-NH2
biotin-GGGGGCꢀAꢀAꢀAFRHDSGYꢀAꢀAꢀAC-NH2
biotin-GGGGGCaaaFRHDSGYaaaC-NH2
biotin-GGGGG[CFRHDSGYC]-NH2
biotin-GGGGG[CAAFRHDSGYAAC]-NH2
biotin-GGGGG[CꢀAꢀAFRHDSGYꢀAꢀAC]-NH2
biotin-GGGGG[CaaFRHDSGYaaC]-NH2
biotin-GGGGG[CAAAFRHDSGYAAAC]-NH2
biotin-GGGGG[CꢀAꢀAꢀAFRHDSGYꢀAꢀAꢀAC]-NH2
biotin-GGGGG[CaaaFRHDSGYaaaC]-NH2
Ac-FRHDSGY-NH2
0.52 ( 0.11
0.71 ( 0.25
0.92 ( 0.22
0.70 ( 0.16
1.14 ( 0.01
0.82 ( 0.20
n.a.h
n.a.
n.a.
n.a.
n.a.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
n.a.
n.a.
2.28 ( 0.01
0.61 ( 0.06
0.62 ( 0.10
1.07 ( 0.39
0.49 ( 0.05
0.52 ( 0.03
1.11 ( 0.29
n.a.
Ac-AAFRHDSGYAA-NH2
Ac-[CFRHDSGYC]-NH2
Ac-[CAAFRHDSGYAAC]-NH2
1205.5/1205.8
1125.4/1125.5
1409.5/1409.7
n.a.
n.a.
n.a.
a Cyclopeptide sequences are indicated by brackets; lower case aa denotes D-amino acids. b Peptides 1-22, RP-HPLC column: Nucleosil 300-7 C18
column. c Peptides 23-26, RP-HPLC column: Discovery BIO Wide Pore C18, Supelco. d The purity of all compounds was >95% according to the HPLC
calculated from AUCs (area under the curves). e [M], peptides 1-22, mass spectrometric analysis was performed with a Bruker APEX II FTICR instrument.
f Peptides 23-26, mass spectrometric analysis was performed with a Bruker Biflex linear TOF mass spectrometer. g The peptide concentration to obtain an
OD ) 1.0 in ELISA experiments. h n.a.: not analyzed.
Figure 1. Outline of the synthesis of cyclic peptides containing the Aꢀ(4–10) epitope (A) and MALDI-FTICR mass spectra of linear peptide (9)
and its cyclic version (16).
is well-known to preferentially stabilize peptides and proteins
in an ordered conformation.40–42 The CD spectra of peptides
3-8 and 17-22 are influenced by the chiral contribution of
aromatic residues (Phe, His, and Tyr) in addition to the CD of
the backbone conformation (Figure 4). The spectra of peptides
5, 8, 19, and 22 also reflect the chiral and conformational effects

1154 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5
Manea et al.
Figure 2. Binding of mouse anti-Aꢀ(1–17) mAb to linear Aꢀ(4–10) containing epitope peptides: (A) comparison of the length of Ala-flanking
regions (peptides 1, 3, and 6); (B) comparison of the type of the Ala-flanking regions (peptides 1, 3, 4, and 5).
Figure 3. Binding of mouse anti-Aꢀ(1–17) mAb to cyclic peptides containing the Aꢀ(4–10) epitope. Comparative studies of the length of Ala-
flanking regions (peptides 16, 17, and 20) (A) or ꢀ-Ala flank regions (peptides 18 and 21) (B) and of the type of the amino acid in the flanks
(peptides 20–22) (C) and (peptides 16-19) (D).
of D-alanine residues. Cyclic disulfides 17-22 are expected to
adopt one major, rigid conformation but the spectral contribution
of the chiral disulfide linkage must be also taken into consid-
eration. The conformational effect of achiral ꢀ-alanine residues
in peptides 4, 7, 18, and 21 is difficult to be evaluated. These
spectra are similar to those of the linear and cyclic compounds
without flanks (data not shown). In consequence, the CD spectra
of the presented peptides (3-8 and 17-22) must be analyzed
with extreme caution.
CD spectra of linear peptides in which the Aꢀ(4–10) epitope
was flanked by R-L-alanine dimers or trimers (3, 6) in water
showed a weak positive band or shoulder near 220 nm and an
intensive negative band at 198–199 nm, indicating the predomi-
nance of unordered structure (Figure 4A).43 Based on their
spectral features, the peptides 3, 6, 17, 20 adopt an ordered,
most probably folded (turn) conformation in TFE (π-π* band
at 207 nm and n-π* band at 222–223 nm) (Figure 4B).43–45
CD spectra of the linear and cyclic peptides in which the
Aꢀ(4–10) epitope were flanked by R-D-alanine dimers or trimers
(5, 8, 19, and 22) showed in water both the chiroptical and
conformational effect of the D-amino acid residues (intensive
positive band near 196–198 nm and a weaker positive band
around 224–226 nm, Figure 4C).46 Compared to the spectra in
water, the CD of peptides 5, 8, 19, and 22 recorded in TFE did
not indicate any significant conformational changes (Figure 4D).
CD spectra of linear and cyclic peptides in which the
Aꢀ(4–10) epitope was flanked by ꢀ-alanine dimers or trimers
(4, 7, 18, and 21) showed in water a positive band near 220
nm, and a negative one around 200 nm (Figure 4E).47 According
to Woody, peptides lacking a discernible regularity in their
conformation might be expected to have weak and variable CD
spectra.45 CD spectra recorded in TFE reflect major conforma-
tional changes; a positive band near 192 nm, a negative band
at about 205 nm and a weak negative shoulder around 220 nm.
All these features indicate the presence of ordered conformer
population(s) (Figure 4F).
NMR Solution Structure. Although the CD spectra recorded
in water indicated the lack of a well defined structure, the

Conformational Flexibility of a Plaque-Specific ꢀ-Amyloid Epitope
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5 1155
Figure 4. CD spectra of linear and cyclic ꢀ-amyloid(4–10)-containing epitope peptides in water and in TFE. Linear peptides flanked by R-L-Ala
dimers (3) or trimers (6) and their cyclic analogues (17 and 20): (A) in water, (B) in TFE. Liner peptides flanked by R-D-Ala dimers (5) or trimers
(8) and their cyclic analogues (19 and 22): (C) in water, (D) in TFE. Linear peptides flanked by ꢀ-Ala dimers (4) or trimers (7) and their cyclic
analogues (18 and 21): (E) in water, (F) in TFE.
peptides might adopt a stable structure that can be defined by
the NMR parameters.48 To investigate this, four peptides (two
linear and two cyclic compounds) were selected for NMR
measurements based on their binding properties (low antige-
nicity: peptides 1 and 16; high antigenicity: peptides 3 and 17).
Therefore, peptides 23-26 were synthesized without the penta-
glycine spacer and biotin that were required for ELISA binding
studies.
Their solution structure was determined by one- and two-
dimensional 1H-1H correlated TOCSY, NOESY, and ROESY
1H NMR measurements.
For all four peptides studied NOESY measurements showed
1
Figure 5. Part of the 2D H-1H TOCSY spectra of peptide 25 with
no cross-peaks at 300 K. Thus, in the case of peptide 25, which
is the smallest cyclic peptide and might have the most rigid
structure, ROESY spectra were collected at 285 K. In this case,
sequential assignment was possible (Figure 5). In the fingerprint
region, only sequential (i, i + 1) NOEs were detected between
backbone NH and the side chain HR. Sequential NH-NH cross
peaks occurred between Phe1-Arg2, Arg2-His3, Asp4-Ser5,
Ser5-Gly6, Tyr7-Cys8. No long-range i to i + 2, or i to i + 3
NOEs were detected. All these data indicate that a high number
60 ms mixing time. Peak assignment in the NH-HR region is shown.
of interchanging conformers are present in solution and it is
not possible to determine a definite solution structure.
The one-dimensional 1H spectra showed only a few overlaps
in the NH region; therefore, direct detection of JNHHR values
was possible. In all cases, these values were between 6 and 7
Hz, in accordance with mostly random coil solution structure.
3

1156 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5
Manea et al.
Scheme 1. Generation of the Conformational Ensemble for Each
Peptide (a-d) and Selection of the Binding Conformation
(e-g)a
Figure 6. Progress of the binding conformation search algorithm. Three
cycles in a schematic 2D conformational space are shown. The fuzzy
gray background represents the visited points by the trajectories of all
molecules. (a) Collect anchor conformations, shown as red dots. These
ACs cannot be closer to each other than F1. Conformational density
values are assigned to each anchor; one representative is shown in blue.
(b) Select the best ACs by q2, where r > 0. The local environment of
the selected ACs is marked in yellow. (c) Remap the selected subspace
with new ACs (green) using the new limit F2. These anchors are also
filtered by their correlation quality. (d) The new subspace is shown in
yellow. (e) This subspace is also populated with much denser ACs. (f)
The AC, selected from the whole set of previously collected ACs is
nominated as the binding conformation candidate of this run which
has the best q2.
a Key: (a) reference conformer generation; (b) molecular dynamics
simulations; (c) trajectories; (d) concatenated trajectory for a single peptide;
(e) concatenated trajectories for all peptides; (f) BC search algorithm; (g)
nomination of prevailing candidates to binding conformations.
Temperature dependence studies at 285, 300, and 310 K of
the chemical shifts for individual NH protons from one-
dimensional spectra were performed for peptide 25. For three
well defined residues Arg2, Ser5, and Tyr7 the calculated ∆δ/
∆T values were -6.7, -6.2, and -5.7 ppb/°C, respectively.
These values might suggest that the backbone amide protons
are somehow protected from exchange with solvent; however,
as the values are closed to -7 ppb/°C, these tend more to be in
random coil.
the whole conformational space of a flexible peptide within the
simulation time.
Computational Methods. The conformational ensembles of
the molecules were obtained computationally (Scheme 1a-d).
To develop starting conformers (RCs, reference conformers) for
the subsequent molecular dynamics simulations a conformational
search24 was applied for each peptide. The conformational
search operates on a stochastic basis by using high energy
molecular dynamics simulations where no solvent effect is taken
into account. Geometry optimization procedures are performed
on the trajectory steps depending on the conformational change.
The RCs are selected from the low energy conformers so that
the conformational diversity between each other should not be
lower than a specified limit. The vector of backbone ꢁ and Ψ
dihedral angles of the CORE defined the conformational space,
while the deviation was computed as the root-mean-square
deviation (rmsd) of the corresponding dihedral vectors. The
search proceeds until no new conformer is found. This type of
conformational analysis was applied both on cyclic and on linear
molecules. The more RCs the molecule has, the more expanded
its accessible conformational space is. Therefore, the number
of RCs can be used as a flexibility measure of the molecule.
MD simulations in aqueous environment and at room
temperature and pressure were launched from all RCs of all
peptides (Scheme 1c). The reasoning behind using more starting
conformers for one peptide is that single MD may not reach
The probabilities of the BCs in the simulated trajectories and
the binding activities have a linear correlation with positive slope
and it enables us to locate regions of BCs. The whole process
of seeking the BC is shown in Scheme 1e-g.
Our novel algorithm24 was utilized to search for these regions
in the CORE space. The CORE space can be represented by
selected dihedrals of the CORE or simply by the Cartesian
coordinates of the CORE atoms. In the first case, the confor-
mational diversity is computed as the rmsd of the dihedral
vectors, while in the second case it is given as the rmsd of the
atomic positions of the superimposed structures.49
The BC search algorithm (Scheme 1f) can locate regions that
have positive Pearson correlation coefficient (r) and high enough
cross-validated r2 (q2) value. The operation of this algorithm,
in an arbitrary, two-dimensional CORE conformational space,
is summarized in Figure 6.
The process starts with collecting anchor conformations (ACs)
from trajectories of all molecules (Scheme 1e). These conforma-
tions cannot be closer to each other than F1, a user defined limit
(Figure 6a). A conformational density array is assigned to each
AC. The number of components of this array is the number of
molecules. The algorithm goes through again on the trajectories,
and develops density values for each AC. The blue circle in

Conformational Flexibility of a Plaque-Specific ꢀ-Amyloid Epitope
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5 1157
Figure 7. Binding conformation of the 14 epitope peptides, superimposed by their CORE, the backbone atoms of the FRHDSGY region. The
amino acids F, R, H, D, S, G, and Y are marked in blue, red yellow, green, orange, cyan, and brown, respectively; only the backbone is shown.
The off-CORE part of the peptides is marked in gray. Conformational density-activity curve for the BC is shown on the right. The horizontal axis
represents the binding activity for each peptide; on the vertical axis is shown the percentage of the given BC in the full conformational ensemble
of the peptide in question.
Table 2. Relative Antibody Binding and the Density of the Binding Conformation in the Full Sampled Space for a Given Peptide
peptide no.
CORE
1/c (OD ) 1) (108 L/mol)
conformational density (%)
20
21
3
17
18
6
4
8
5
19
1
[CAAA
[CꢀAꢀAꢀA
AA
[CAA
[CꢀAꢀA
AAA
ꢀAꢀA
aaa
aa
[Caa
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
AAAC]
ꢀAꢀAꢀAC]
AA
AAC]
ꢀAꢀAC]
AAA
ꢀAꢀA
aaa
aa
aaC]
2.04
1.92
1.92
1.64
1.61
1.43
1.41
1.22
1.09
0.93
0.91
0.90
0.88
0.44
0.69
0.56
0.64
0.65
0.72
0.68
0.55
0.61
0.38
0.19
0.20
0.09
0.19
0.00
22
7
16
[Caaa
ꢀAꢀAꢀA
[C
aaaC]
ꢀAꢀAꢀA
C]
Figure 6a symbolizes the threshold, δ, for the density calculation,
which is kept constant during the whole run. The threshold was
applied in the current study as a simple cutoff, however,
Gaussian weighting is optionally available in our implementa-
tion. The q2 and r values are evaluated for all ACs using their
density arrays and the binding activities. Those ACs are selected
that has positive r and high enough q2 value.
The region that is closer to the selected best ACs than F1
(Figure 6b) is remapped with new, much densely populated ACs
with a lower diversity limit, F2 (Figure 6c). Figure 6d shows
the ACs selected in this cycle and their environment to be
remapped in the next turn (Figure 6e). The algorithm proceeds
stepwise at decreasing F values until the repeat count reaches
the user defined limit (NF). As the result of this type of
calculation may depend on the sequence of conformations in
the input trajectories (Scheme 1e), this algorithm is repeated
Nrand times on randomized sequences.
of linear and cyclic peptides containing the plaque-specific
ꢀ-amyloid(4–10) epitope was investigated. Cyclization of the
native Aꢀ(4–10) epitope resulted in substantially reduced
antibody binding compared to its linear version. However, the
unfavorable effect of cyclization was completely compensated
by insertion of R- or ꢀ-alanine dimer or trimer at both N- and
C-terminus of the epitope in the respective cyclopeptides.
Furthermore, we found that the chirality of R-Ala residues had
a marked influence on the antibody recognition. Cyclopeptides
comprising the epitope flanked by R-D-alanine had the lowest
relative binding (compared to R-L-alanine and ꢀ-alanine flanking
regions); however, higher than the cyclic peptide without any
flanks.
CD spectroscopy was applied to analyze the conformation
of the epitope peptides. CD spectra of cyclic peptides in which
the Aꢀ(4–10) epitope was flanked by R-L-alanine dimers or
trimers, recorded in water, indicated the presence of some
ordered conformers as compared to those of the corresponding
linear peptides. No significant effect of cyclization on the
solution conformation of the peptides containing the epitope
with R-D-alanine or ꢀ-alanine flanks was determined in water.
Linear and cyclic peptides comprising R-L-alanine or ꢀ-alanine
dimer and trimer flanks exhibited some ordered structure in TFE.
In contrast, the conformation of linear and cyclic peptides with
R-D-alanine flanks did not show significant changes in TFE.
Those ACs that has high enough q2 and positive r values are
clustered50,51 according to their conformational diversity (Scheme
1g). The cluster, which has the best q2 value, is accepted as the
BCC. Two clusters were found, with q2 0.64 and 0.67 and with
r 0.88 for both. The latter was accepted as the binding
conformation (Figure 7, Table 2). The significance of the model
was validated via F-test (F ) 31.09).
Conclusion
A linear relationship between the binding activity and the
probability of the binding conformation in the solvent for a set
of molecules allowed us to develop a novel 3D QSAR method
In the present study, the effect of non-native flanking
sequences on antibody binding and conformational flexibility

1158 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5
Manea et al.
Germany). The following side-chain protections of amino acid
derivatives were employed: Fmoc-Tyr(tBu)-OH, Fmoc-Ser(tBu)-
OH, Fmoc-Asp(OtBu)-OH, Fmoc-His(Trt)-OH and Fmoc-Arg(Pbf)-
OH. The synthetic protocol was as follows: (i) DMF washing (3 ×
1 min); (ii) Fmoc deprotection for 15 min using 2% DBU and 2%
piperidine in DMF; (iii) DMF washing (6 × 1 min); (iv) coupling
of 5 equiv of Fmoc amino acid/ PyBOP/NMM in DMF for 45 min;
(v) DMF washing (3 × 1 min). After completion of the synthesis,
the N-terminus was biotinylated using 5 equiv of D-(+)-Biotin/
PyBOP/ NMM. The peptides were then cleaved from the resin at
25 °C for 3 h using a mixture of TFA, triethylsilane, and deionized
water (95: 2.5: 2.5, v/v/v). The crude products were precipitated
with cold tert-butyl methyl ether, washed three times with diethyl
ether and solubilized in 5% aqueous acetic acid prior to freeze-
drying. The crude peptides were purified by reverse phase-high
performance liquid chromatography (RP-HPLC) on a preparative
C18 column. Purified peptides were analyzed by analytical RP-
HPLC, and peptide structures characterized by matrix assisted laser
desorption ionization-Fourier transform ion cyclotron resonance
mass spectrometry (MALDI-FTICR MS) (Table 1).
Synthesis of Biotinylated Cyclic Aꢀ(4–10) Epitope Peptides.
Linear precursor peptides used for cyclization (biotin-GGGGGC-
FRHDSGYC-NH2 (9), biotin-GGGGGCAAFRHDSGYAAC-NH2
(10) biotin-GGGGGCꢀAꢀAFRHDSGYꢀAꢀAC-NH2 (11), biotin-
GGGGGCaaFRHDSGYaaC-NH2 (12), biotin-GGGGGCAAAFRH-
DSGYAAAC-NH2 (13), biotin-GGGGGCꢀAꢀAꢀAFRHDSGY-
ꢀAꢀAꢀAC-NH2 (14), and biotin-GGGGGCaaaFRHDSGYaaaC-
NH2 (15)) were synthesized by SPPS using Fmoc/tBu chemistry
according to the protocol described above. Trityl was used for the
side-chain protection of cysteine. Prior to cyclization, the crude
linear peptides were purified by preparative RP-HPLC, and purified
peptides were characterized by analytical RP-HPLC and MALDI-
FTICR mass spectrometry (Table 1). All cyclization reactions were
carried out by air oxidation in solution. Linear peptides containing
two Cys residues were dissolved in a 1:1 (v/v) mixture of 50 mM
aqueous ammonium acetate (pH 6) and dimethyl sulfoxide at a
peptide concentration of 0.2 mg/mL, and then the reaction mixtures
were stirred at 25 °C for 24 h. After freeze-drying, the crude
peptides (16-22) were purified by RP-HPLC on a preparative C18
column and characterized by analytical RP-HPLC and MALDI-
FTICR MS. Retention times and mass spectrometric data are
presented in Table 1.
Synthesis of Linear and Cyclic Aꢀ(4–10) Epitope Peptides
for NMR Studies. Linear and cyclic peptides Ac-FRHDSGY-NH2
(23), Ac-AAFRHDSGYAA-NH2 (24), Ac-[CFRHDSGYC]-NH2
(25), and Ac-[CAAFRHDSGYAAC]-NH2 (26) were synthesized
as described above with minor modifications. After completion of
the synthesis, the N-terminus was acetylated using a mixture of
Ac2O (1 mL), DIEA (1 mL) in DMF (3 mL) for 30 min at room
temperature. The peptides were characterized by analytical RP-
HPLC and MALDI-TOF mass spectrometry (Table 1).
High-Performance Liquid Chromatography (HPLC). Ana-
lytical RP-HPLC was performed on a Bio-Rad instrument (Bio-
Rad Laboratories, Richmond, CA) using an analytical Nucleosil
300-7 C18 column (250 × 4 mm, 300 Å, 7 µm; Macherey-Nagel,
Dueren, Germany) or Discovery BIO Wide Pore C18, Supelco (15
cm × 4.6 mm, 3 µm) as a stationary phase. Linear gradient elution
(0 min 0% B; 5 min 0% B; 50 min 90% B) with eluent A (0.1%
TFA in water) and eluent B (0.1% TFA in acetonitrile-water,
80:20, v/v) was used at a flow rate of 1 mL/min. Samples were
dissolved in eluent A and the peptides detected at 220 nm.
Preparative purifications of linear and cyclic peptides were carried
out on a Knauer HPLC system (Knauer, Berlin, Germany) using a
preparative C18 column (GROM-SIL 120 ODS-4 HE, 10 µm, 250
× 20 mm, pore size 120 Å; Grom, Herrenberg-Kayh, Germany).
Linear gradient elution (0 min 10% B; 5 min 10% B; 65 min 75%
B) was employed using the same mobile phases as described above,
with a flow rate of 10 mL/min. Peptides were dissolved in solvent
A, and the peak detection was performed at 220 nm.
exclusively for flexible molecules. The algorithm was coded in
JAVA language using standard ChemAxon tools. This type of
3D QSAR application avoids using grids and multidimensional
regression methods. We accepted the best model, with q2 )
0.67, as the binding conformation obtained from the calculations.
According to this model the backbone of the examined
FRHDSGY sequence adopts an extended conformation, with a
turn like structure around serine. It has to be noted that other
model candidates that had unacceptable q2 values mostly
adopted such globular conformations, where some of the side
chains of the CORE amino acids were buried. On the other hand,
in the case of the present accepted model the side chains are
accessible to the target protein (antibody).
The conformational analysis suggests that the peptides with
R-D-Ala flanks have the less flexible core epitope regions and
the ꢀ-Ala flanks result in increased flexibility. Comparing the
antibody recognition of the peptides in which the epitope was
flanked by R-L-Ala and ꢀ-Ala residues, the cyclic peptides have
slightly increased antibody binding compared with the linear
ones. However, the differences are not significant except for
the case of linear and cyclic epitope peptides containing three
ꢀ-Ala in flanking regions. As presented in Table 2, according
to the molecular dynamics simulations, the linear peptide has
fairly low density of binding conformation compared with the
cyclic analogue. In the case of the peptides with R-D-Ala flanks
or without flanks, the cyclic compounds have lower antibody
binding compared to the linear ones, the cyclization of these
peptides resulting in less favored and fairly rigid conformations.
The solution structure determined by NMR spectroscopy
indicated that a high number of interchanging conformers were
present in solution and it was not possible to determine a definite
solution structure of the investigated peptides. The lack of NOE
peaks means that the distance between the H atoms is higher
than 6 Å. Based on the modeling results, all i and i + 2, i + 3
distances were analyzed and the values were higher than 6 Å.
Moreover, sequential distances below 6 Å were those detected
by NMR.
Taken together, these results indicate that by appropriate
combination of cyclization and insertion of Ala-based flanking
regions it is feasible to construct simple B-cell epitope peptides
with improved antigenicity.
Experimental Section
Materials. All amino acid derivatives and NovaSyn TGR resin
were purchased from NovaBiochem (Läufelfingen, Switzerland) and
D-(+)-Biotin from Calbiochem (Darmstadt, Germany). Scavengers,
coupling agents, and cleavage reagents (triethylsilane, 4-methyl-
morpholine (NMM), piperidine, 1,8-diazabicyclo[5.4.0]undec-7-ene
(DBU), trifluoroacetic acid (TFA), diisopropylethylamine (DIEA),
acetic anhydride (Ac2O)) were obtained from Fluka (Buchs,
Switzerland), benzotriazole-1-yloxytrispyrrolidinophosphonium
hexafluorophosphate (PyBOP) was from NovaBiochem. Dimeth-
ylformamide (DMF) was purchased from Acros Organics (Geel,
Belgium), while ethanol, tert-butyl methyl ether, and acetonitrile
were from Riedel-deHäen (Seelze, Germany). Other reagents and
solvents were of analytical grade or highest available purity.
Synthesis of Biotinylated Linear Aꢀ(4–10) Epitope Peptides.
Peptides, biotin-GGGGGFRHDSGY-NH2 (1), biotin-GGGGGYGS-
DHRF-NH2 (2), biotin-GGGGGAAFRHDSGYAA-NH2 (3), biotin-
GGGGGꢀAꢀAFRHDSGYꢀAꢀA-NH2 (4), biotin-GGGGGaaFRHDS-
GYaa-NH2(5),biotin-GGGGGAAAFRHDSGYAAA-NH2(6),biotin-
GGGGGꢀAꢀAꢀAFRHDSGYꢀAꢀAꢀA-NH2 (7), and biotin-
GGGGGaaaFRHDSGYaaa-NH2 (8) were synthesized on a NovaSyn
TGR resin (0.23 mmol/g coupling capacity) by 9-fluorenyl-
methoxycarbonyl/tert-butyl (Fmoc/tBu) chemistry, using a semi-
automated Peptide Synthesizer EPS-221 (ABIMED, Langenfeld,
Mass Spectrometry (MS). MALDI-FTICR mass spectrometric
analysis was performed with a Bruker APEX II FTICR instrument

Conformational Flexibility of a Plaque-Specific ꢀ-Amyloid Epitope
Table 3. Peptides Which Were Subjected to QSAR Analysisa
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5 1159
peptide no.
CORE
tvq (ns)
NRC
Nwat
tsol (ns)
20
21
3
17
18
6
4
8
5
19
1
[CAAA
[CꢀAꢀAꢀA
AA
[CAA
[CꢀAꢀA
AAA
ꢀAꢀA
aaa
aa
[Caa
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
FRHDSGY
AAAC]
ꢀAꢀAꢀAC]
AA
AAC]
ꢀAꢀAC]
AAA
ꢀAꢀA
aaa
aa
aaC]
11.6692
197.992
12.3649
15.9262
35.4539
24.0858
26.9148
5.42628
6.5925
9.35551
10.4792
4.0661
71.3566
21.6124
142
848
74
117
306
117
194
68
57
81
51
61
1461
1412
1315
1415
1388
1424
1345
1417
1300
1444
1221
1517
1408
1234
28.4
169.6
14.8
23.4
61.2
23.4
38.8
13.6
11.4
16.2
10.2
12.2
109
22
7
16
[Caaa
ꢀAꢀAꢀA
[C
aaaC]
ꢀAꢀAꢀA
C]
545
75
15
a The simulation time required by the conformational search (tvq), the number of low energy conformers (NRC), the average number of water molecules
(Nwat) added to each peptide, and trajectory length subjected to data mining (tsol) are shown. The peptides in the table are listed according to the increase of
the antibody binding.
equipped with an actively shielded 7T superconducting magnet, a
cylindrical infinity ICR analyzer cell, and an external Scout 100
fully automated X-Y target stage MALDI source with pulsed
collision gas (Bruker Daltonics, Bremen, Germany). The pulsed
MALDI nitrogen laser was operated at 337 nm. A 100 µg/µL
solution of 2,5-dihydroxybenzoic acid (DHB, Aldrich, Steinheim,
Germany) in acetonitrile: 0.1% trifluoroacetic acid in water (2:1,
v/v) was used as the matrix. Aliquots of 0.5 µL matrix solution
and 0.5 µL of sample solution were mixed on the stainless steel
MALDI sample target and allowed to dry. Calibration was
performed with a standard peptide mixture within an m/z range of
approximately 4000.
ammonium-d-camphor-10-sulfonate (Katayama Chemical, Japan)
in water. Distilled water and trifluoroethanol (TFE) (Fluka, Buchs,
Switzerland) were used as solvents. The concentration of the
samples was 200 µM. Spectra were averaged over six scans between
190 and 260 nm. Results are expressed in terms of mean residue
ellipticities ([Θ] in deg cm2 dmol-1) after subtraction of the solvent
baseline.
NMR Spectroscopy. Measurements were performed on a Bruker
DRX 500 MHz spectrometer equipped with a z-gradient 5 mm triple
resonance probe head. Temperature values were checked with the
1,2-ethanediol calibration method. Peptides were dissolved in H2O,
and samples contained 10% D2O. Chemical shifts were referenced
to 2,2 dimethyl-2-silapentane-5-sulfonate (DSS) as internal standard.
Two dimensional 1H-1H correlation spectra: TOCSY, NOESY,
and ROESY were measured by using standard Bruker pulse
programs. Spectra were analyzed by the Sparky software.
Computational Details. Reference conformers (RCs) were
collected via high energy molecular dynamics simulations in
vacuum. Gromos 43b1 force field was used with the simulation
temperature set to 600 K, Berendsen type temperature coupling52
with τt ) 0.1 and 1 fs time step. All bond lengths were
constrained.53 The default values were used for other MD param-
eters as implemented in Gromacs 3.3.1.54–56 Table 3 shows the
simulation time required by the conformational search (tvq), and
MALDI-TOF MS was carried out with a Bruker Biflex linear
TOF mass spectrometer (Bruker Daltonics, Bremen, Germany). A
saturated solution of R-cyano-4-hydroxy-cinnamic acid (HCCA)
in acetonitrile/0.1% trifluoroacetic acid in water (2:1 v/v) was used
as the matrix. Acquisition of spectra was carried out at an
acceleration voltage of 20 kV and a detector voltage of 1.5 kV.
Enzyme-Linked Immunosorbent Assay (ELISA). Indirect
ELISA analyses were performed using a standard dilution of the
mouse anti-Aꢀ(1–17) monoclonal antibody (mAb) (clone 6E10,
Chemicon, Temecula, CA) in combination with nine serial dilutions
of the peptides. The dilutions were made in PBS solution containing
5 mM Na2HPO4 and 150 mM NaCl, pH ) 7.5. Ninety-six-well
ELISA plates (BioRad, Hercules, CA) were coated with 100 µL/
well streptavidin solution (Sigma-Aldrich, Steinheim, Germany) in
PBS (2.5 µg/mL) at 25 °C for 2 h. After the wells were washed
with PBS-T (0.05% Tween-20 in PBS, pH 7.5), 100 µL/well of
biotinylated epitope peptides was added and the mixtures were
incubated at 25 °C for 2 h. The plates were washed four times
with 200 µL/well PBS-T, and the nonspecific adsorption sites were
blocked with 5% BSA (w/v) in PBS, by incubation at 4 °C for
12 h. Then, 100 µL/well of the mouse anti-Aꢀ(1–17) monoclonal
antibody (1:4000 dilution in 5% BSA) was added to each well.
The plates were incubated at 25 °C for 2 h and then washed four
times with washing buffer. A 100 µL solution of peroxidase goat
antimouse IgG (Jackson Immuno Research, West Grove, PA),
diluted 5000 times in 5% BSA, was added to each well and the
plates incubated at 25 °C for two hours, then washed three times
with washing buffer and two times with 0.05 M sodium
phosphate-citrate buffer, pH 5. A 100 µL volume of o-phenylene-
diamine dihydrochloride (OPD) (Merck, Darmstadt, Germany) in
phosphate-citrate substrate buffer at 1 mg/mL containing 2 µL of
30% hydrogen-peroxide (Merck, Darmstadt, Germany) per 10 mL
was added, and absorbance was measured at 450 nm on a Wallac
1420 Victor2 ELISA Plate Counter (PerkinElmer, Boston, MA).
The concentration of the peptide solution, which gave an OD450 of
1, was calculated, and data are presented in Table 1.
the number of the obtained low energy reference conformers (NRC
)
for the peptides which were subjected to QSAR analysis.
RCs were surrounded by water molecules, in the framework of
the SPC model, and Na+ and Cl- counterions were added to the
side chains of His and Arg, respectively, to neutralize their charge
at pH ) 7. The average number of water molecules (Nwat) added
to each peptide is shown in Table 3. To equilibrate the solvent
molecules, the peptide and the ions were kept frozen57 in the
subsequent MD simulations, launched for all RCs, with the
temperature set to 300 K, time step to 2 fs and the van der Waals
and Coulomb cut offs to 1 nm for a 10 ps simulation. Then, the
solute and the ions were released, and simulations were carried
out for 2 × 0.2 ns simulation time. The first 0.2 ns were used to
equilibrate the peptide structure in the solvent environment and were
excluded from further analysis. Data mining was performed on the
second 0.2 ns long simulation period for every RCs. The total
trajectory length subjected to data mining for each peptide is shown
(tsol) in Table 3.
The BC search algorithm was written in pure java and the
standard ChemAxon tools58 were used, the trajectories and
the anchor conformations were stored in a MySQL database. The
CORE space for these 14 peptides was defined a priori as the
backbone ꢁ and Ψ dihedrals of the amino acid sequence
FRHDSGY. The rmsd of the dihedral vectors was used as a
diversity measure.
Circular Dichroism Spectroscopy (CD). Circular dichroism
spectra were recorded on a JASCO spectropolarimeter, model J-715,
at 20 °C in quartz cells of 0.05 cm path length, under constant
nitrogen flush. The instrument was calibrated with 0.06% (w/v)
It has to be noted that the BC search using rmsd of the
superimposed atomic coordinates of the CORE was beyond our

1160 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5
Manea et al.
(8) Drakopoulou, E.; Uray, K.; Mezö, G.; Price, M. R.; Vita, C.; Hudecz,
F. Synthesis and antibody recognition of mucin 1 (MUC1)-alpha-
conotoxin chimera. J. Peptide Sci. 2000, 6, 175–185.
(9) Hudecz, F. Manipulation of epitope function by modification of peptide
structure: A minireview. Biologicals 2001, 29, 197–207.
(10) Mezö, G.; Drakopoulou, E.; Paál, V.; Rajnavölgyi, E.; Vita, C.; Hudecz,
F. Synthesis and immunological studies of alpha-conotoxin chimera
containing an immunodominant epitope from the 268–284 region of
HSV gD protein. J. Peptide Res. 2000, 55, 7–17.
(11) Tugyi, R.; Mezö, G.; Fellinger, E.; Andreu, D.; Hudecz, F. The effect
of cyclization on the enzymatic degradation of herpes simplex virus
glycoprotein D derived epitope peptide. J. Peptide Sci. 2005, 11, 642–
649.
computational capacity. From the same reason simple density cut
off was utilized during the calculations, and no Gaussian weighting
was applied.
After scanning24 the parameter space of the algorithm the
following input values were utilized: the number of cycles per run
(NF) was set to 5 and in each cycle the best 20% of the ACs was
accepted. The F values were scaled down by 0.9 in each cycle.
The initial F value, the F1, was set equal to δ. The δ was set to 45°
dihedral angle rmsd. The number of runs launched from randomized
sequence of conformations the Nrand value was set to 20. The ACs
which had higher q2 value than 0.6 were clustered.
(12) Tugyi, R.; Uray, K.; Iván, D.; Fellinger, E.; Perkins, A.; Hudecz, F.
Partial D-amino acid substitution: Improved enzymatic stability and
preserved Ab recognition of a MUC2 epitope peptide. Proc. Natl.
Acad. Sci. U.S.A. 2005, 102, 413–418.
(13) Wilkinson, K. A.; Vordermeier, M. H.; Kajtár, J.; Jurcevic, S.;
Wilkinson, R.; Ivanyi, J.; Hudecz, F. Modulation of peptide specific
T cell responses by non-native flanking regions. Mol. Immunol. 1997,
34, 1237–1246.
(14) Mezö, G.; de Oliveira, E.; Krikorian, D.; Feijlbrief, M.; Jakab, A.;
Tsikaris, V.; Sakarellos, C.; Welling-Wester, S.; Andreu, D.; Hudecz,
F. Synthesis and comparison of antibody recognition of conjugates
containing herpes simplex virus type 1 glycoprotein D epitope VII.
Bioconjugate Chem. 2003, 14, 1260–1269.
(15) Manea, M.; Mezö, G.; Hudecz, F.; Przybylski, M. Polypeptide
conjugates comprising a beta-amyloid plaque-specific epitope as new
vaccine structures against Alzheimer’s disease. Biopolymers 2004, 76,
503–511.
Acknowledgment. This work has been partially supported
by grants from the Deutsche Forschungsgemeinschaft and the
EU network “Ligand Binders to the Human Proteome”. The
expert assistance by Dmitry Galetskiy and Reinhold Weber with
FTICR mass spectrometery is gratefully acknowledged. The
QSAR development was funded by the Agency for Research
Fund Management and Research Exploitation: GVOP-AKF-
2004-3.1.1-0351. We acknowledge the support of the InfoPark
Foundation, ChemAxon Ltd., Richter Ltd., and Gaussian, Inc.
This work was also supported by GVOP-3.2.1.-2004-04-0005/
3.0 and by the Hungarian National Science Fund (OTKA T
03456, D 48459).
Supporting Information Available: HPLC chromatograms and
mass spectra of the compounds and temperature dependence of the
1H chemical shift for three selected residues. This material is
available free of charge via the Internet at http://pubs.acs.org.
(16) Manea, M.; Hudecz, F.; Przybylski, M.; Mezö, G. Synthesis, solution
conformation, and antibody recognition of oligotuftsin-based conju-
gates containing a beta-amyloid(4–10) plaque-specific epitope. Bio-
conjugate Chem. 2005, 16, 921–928.
(17) Hopfinger, A. J.; Wang, S.; Tokarski, J. S.; Jin, B. Q.; Albuquerque,
M.; Madhav, P. J.; Duraiswami, C. Construction of 3D-QSAR models
using the 4D-QSAR analysis formalism. J. Am. Chem. Soc. 1997, 119,
10509–10524.
References
(1) Weiner, H. L.; Lemere, C. A.; Maron, R.; Spooner, E. T.; Grenfell,
T. J.; Mori, C.; Issazdeh, S.; Hancock, W. W.; Selkoe, D. J. Nasal
administration of amyloid-beta peptide decreases cerebral amyloid
burden in a mouse model of Alzheimer’s disease. Ann. Neurol. 2000,
48, 567–579.
(18) Vedani, A.; Briem, K.; Dobler, M.; Dollinger, H.; McMasters, D. R.
Multiple-conformation and protonation-state representation in 4D-
QSAR: The neurokinin-1 receptor system. J. Med. Chem. 2000, 43,
4416–4427.
(19) Martinek, T. A.; Ötvös, F.; Dervarics, M.; Tóth, G.; Fülöp, F. Ligand-
based prediction of active conformation by 3D-QSAR flexibility
descriptors and their application in 3 + 3D-QSAR models. J. Med.
Chem. 2005, 48, 3239–3250.
(20) Bhonsle, J. B.; Wang, Z. X.; Tamamura, H.; Fujii, N.; Peiper, S. C.;
Trent, J. O. A simple, automated quasi-4D-QSAR, quasi-multi way
PLS approach to develop highly predictive QSAR models for highly
flexible CXCR4 inhibitor cyclic pentapeptide ligands using scripted
common molecular modeling tools. QSAR Comb. Sci. 2005, 24, 620–
630.
(21) Bhonsle, J. B.; Bhattacharjee, A. K.; Gupta, R. K. Novel semi-
automated methodology for developing highly predictive QSAR
models: application for development of QSAR models for insect
repellent amides. J. Mol. Model. 2007, 13, 179–208.
(22) Dervarics, M.; Otvos, F.; Martinek, T. A. Development of a chirality-
sensitive flexibility descriptor for 3 + 3D-QSAR. J. Chem. Inf. Mod.
2006, 46, 1431–1438.
(23) Kalászi, A.; Farkas, Ö. Lead conformer prediction based on a library
of flexible molecules. THEOCHEM 2003, 666, 645–649.
(24) Kalászi, A.; Imre, G.; Jákli, I.; Farkas, Ö. Identification of the bioactive
conformation for mucin epitope peptides. THEOCHEM 2007, 823,
16–27.
(2) Schenk, D.; Barbour, R.; Dunn, W.; Gordon, G.; Grajeda, H.; Guido,
T.; Hu, K.; Huang, J. P.; Johnson-Wood, K.; Khan, K.; Kholodenko,
D.; Lee, M.; Liao, Z. M.; Lieberburg, I.; Motter, R.; Mutter, L.;
Soriano, F.; Shopp, G.; Vasquez, N.; Vandevert, C.; Walker, S.;
Wogulis, M.; Yednock, T.; Games, D.; Seubert, P. Immunization with
amyloid-beta attenuates Alzheimer disease-like pathology in the
PDAPP mouse. Nature 1999, 400, 173–177.
(3) Janus, C.; Pearson, J.; McLaurin, J.; Mathews, P. M.; Jiang, Y.;
Schmidt, S. D.; Chishti, M. A.; Horne, P.; Heslin, D.; French, J.;
Mount, H. T. J.; Nixon, R. A.; Mercken, M.; Bergeron, C.; Fraser,
P. E.; St. George-Hyslop, P.; Westaway, D. A beta peptide immuniza-
tion reduces behavioural impairment and plaques in a model of
Alzheimer’s disease. Nature 2000, 408, 979–982.
(4) Bard, F.; Cannon, C.; Barbour, R.; Burke, R. L.; Games, D.; Grajeda,
H.; Guido, T.; Hu, K.; Huang, J. P.; Johnson-Wood, K.; Khan, K.;
Kholodenko, D.; Lee, M.; Lieberburg, I.; Motter, R.; Nguyen, M.;
Soriano, F.; Vasquez, N.; Weiss, K.; Welch, B.; Seubert, P.; Schenk,
D.; Yednock, T. Peripherally administered antibodies against amyloid
beta-peptide enter the central nervous system and reduce pathology
in a mouse model of Alzheimer disease. Nature Med. 2000, 6, 916–
919.
(25) Janssen, L. H. M. Conformational flexibility and receptor interaction.
(5) DeMattos, R. B.; Bales, K. R.; Cummins, D. J.; Dodart, J. C.; Paul,
S. M.; Holtzman, D. M. Peripheral anti-A beta antibody alters CNS
and plasma A beta clearance and decreases brain A beta burden in a
mouse model of Alzheimer’s disease. Proc. Natl. Acad. Sci. U.S.A.
2001, 98, 8850–8855.
Bioorg. Med. Chem. 1998, 6, 785–788.
(26) Yung-Chi, Cheng.; Prusoff, W. H. Relationship between the inhibition
constant (Ki) and the concentration of inhibitor which causes 50%
inhibition (I50) of an enzymatic reaction. Biochem. Pharmacol. 1973,
22, 3099–3108.
(6) Morgan, D.; Diamond, D. M.; Gottschall, P. E.; Ugen, K. E.; Dickey,
C.; Hardy, J.; Duff, K.; Jantzen, P.; DiCarlo, G.; Wilcock, D.; Connor,
K.; Hatcher, J.; Hope, C.; Gordon, M.; Arendash, G. W. A beta peptide
vaccination prevents memory loss in an animal model of Alzheimer’s
disease. Nature 2000, 408, 982–985.
(27) Nicklaus, M. C.; Wang, S. M.; Driscoll, J. S.; Milne, G. W. A.
Conformational-changes of small molecules binding to proteins.
Bioorg. Med. Chem. 1995, 3, 411–428.
(28) Perola, E.; Charifson, P. S. Conformational analysis of drug-like
molecules bound to proteins: An extensive study of ligand reorganiza-
tion upon binding. J. Med. Chem. 2004, 47, 2499–2510.
(29) Stockwell, G. R.; Thornton, J. M. Conformational diversity of ligands
bound to proteins. J. Mol. Biol. 2006, 356, 928–944.
(30) Steer, D.; Lew, R.; Perlmutter, P.; Smith, A. I.; Aguilar, M. I. Inhibitors
of metalloendopeptidase EC 3.4.24.15 and EC 3.4.24.16 stabilized
against proteolysis by the incorporation of beta-amino acids. Bio-
chemistry 2002, 41, 10819–10826.
(7) McLaurin, J.; Cecal, R.; Kierstead, M. E.; Tian, X.; Phinney, A. L.;
Manea, M.; French, J. E.; Lambermon, M. H. L.; Darabie, A. A.;
Brown, M. E.; Janus, C.; Chishti, M. A.; Horne, P.; Westaway, D.;
Fraser, P. E.; Mount, H. T. J.; Przybylski, M.; St George-Hyslop, P.
Therapeutically effective antibodies against amyloid-beta peptide target
amyloid-beta residues 4–10 and inhibit cytotoxicity and fibrillogenesis.
Nature Med. 2002, 8, 1263–1269.

Conformational Flexibility of a Plaque-Specific ꢀ-Amyloid Epitope
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 5 1161
(31) Briand, J. P.; Benkirane, N.; Guichard, G.; Newman, J. F. E.;
vanRegenmortel, M. H. V.; Brown, F.; Muller, S. A retro-inverso
peptide corresponding to the GH loop of foot-and-mouth disease virus
elicits high levels of long-lasting protective neutralizing antibodies.
Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 12545–12550.
(32) Hamamoto, K.; Kida, Y.; Zhang, Y.; Shimizu, T.; Kuwano, K.
Antimicrobial activity and stability to proteolysis of small linear
cationic peptides with D-amino acid substitutions. Microbiol. Immunol.
2002, 46, 741–749.
(33) Conley, A. J.; Conard, P.; Bondy, S.; Dolan, C. A.; Hannah, J.; Leanza,
W. J.; Marburg, S.; Rivetna, M.; Rusiecki, V. K.; Sugg, E. E.;
Vanmiddlesworth, F.; Warne, S. A.; Ulrich, J. T.; Rudbach, J. A.;
Tolman, R. L.; Emini, E. A. Immunogenicity of Synthetic Hiv-1 Gp120
V3-Loop Peptide-Conjugate Immunogens. Vaccine 1994, 12, 445–
451.
(34) Muller, S.; Plaue, S.; Samama, J. P.; Valette, M.; Briand, J. P.; van
Regenmortel, M. H. V. Antigenic properties and protective capacity
of a cyclic peptide corresponding to site-A of Influenza-virus hemag-
glutinin. Vaccine 1990, 8, 308–314.
(35) Schulzegahmen, U.; Klenk, H. D.; Beyreuther, K. Immunogenicity
of loop-structured short synthetic peptides mimicking the antigenic
site A of Influenza-virus hemagglutinin. Eur. J. Biochem. 1986, 159,
283–289.
(36) Williams, W. V.; Kieberemmons, T.; Vonfeldt, J.; Greene, M. I.;
Weiner, D. B. Design of bioactive peptides based on antibody
hypervariable region structures - development of conformationally
constrained and dimeric peptides with enhanced affinity. J. Biol. Chem.
1991, 266, 5182–5190.
(37) Schlosser, G.; Mezö, G.; Kiss, R.; Vass, E.; Majer, Z.; Feijlbrief, M.;
Perczel, A.; Bösze, S.; Welling-Wester, S.; Hudecz, F. Synthesis,
solution structure analysis and antibody binding of cyclic epitope
peptides from glycoprotein D of Herpes simplex virus type I. Biophys.
Chem. 2003, 106, 155–171.
(42) Nelson, J. W.; Kallenbach, N. R. Stabilization of the ribonuclease
S-peptide R helix by trifluoroethanol. Proteins 1986, 1, 211–217.
(43) Woody, R. W. In Circular dichroism - Principles and applications;
Woody, R. W., Nakanishi, K., Berova, N., Eds.; VCH Publishers, Inc.:
New York, 1996; pp 473–496.
(44) Vass, E.; Hollósi, M.; Besson, F.; Buchet, R. Vibrational spectroscopic
detection of beta- and gamma-turns in synthetic and natural peptides
and proteins. Chem. ReV. 2003, 103, 1917–1954.
(45) Woody, R. W. In The peptides-Analysis, Synthesis, Biology; Uden-
friend, S. Meienhofer, J., Eds.; Academic Press, Inc.: New York, 1985;
pp 15–114.
(46) Uray, K.; Kajtár, J.; Vass, E.; Price, M. R.; Hollósi, M.; Hudecz, F.
Effect of D-amino acid substitution in a mucin 2 epitope on mucin-
specific monoclonal antibody recognition. Arch. Biochem. Biophys.
2000, 378, 25–32.
(47) Cheng, R. P.; Gellman, S. H.; DeGrado, W. F. beta-peptides: From
structure to function. Chem. ReV. 2001, 101, 3219–3232.
(48) Iverson, H. A.; Fox, D., III; Nadler, L. S.; Klevit, R. E.; Nathanson,
N. M. Identification and structural determination of the M3 muscarinic
acetylcholine receptor baselatoral sorting signal. J. Biol. Chem. 2005,
280, 24568–24575.
(49) Heisterberg, D. J. Quaternion fit source code. http://www.ccl.net/
cca/software/SOURCES/C/quaternion-mol-fit/quatfit. 1990.
(50) Barnard, J. M.; Downs, G. M. Chemical fragment generation and
clustering software. J. Chem. Inf. Comput. Sci. 1997, 37, 141–142.
(51) Brown, R. D.; Martin, Y. C. Use of structure Activity data to compare
structure-based clustering methods and descriptors for use in compound
selection. J. Chem. Inf. Comput. Sci. 1996, 36, 572–584.
(52) Berendsen, H. J. C.; Postma, J. P. M.; Vangunsteren, W. F.; Dinola,
A.; Haak, J. R. Molecular-dynamics with coupling to an external bath.
J. Chem. Phys. 1984, 81, 3684–3690.
(53) Hess, B.; Bekker, H.; Berendsen, H. J. C.; Fraaije, J. G. E. M. LINCS:
A linear constraint solver for molecular simulations. J. Comput. Chem.
1997, 18, 1463–1472.
(54) Berendsen, H. J. C.; Vanderspoel, D.; Vandrunen, R. Gromacs - A
message-passing parallel molecular-dynamics implementation. Comput.
Phys. Commun. 1995, 91, 43–56.
(55) Lindahl, E.; Hess, B.; van der Spoel, D. GROMACS 3.0: a package
for molecular simulation and trajectory analysis. J. Mol. Modeling
2001, 7, 306–317.
(56) van der Spoel, D.; Lindahl, E.; Hess, B.; Groenhof, G.; Mark, A. E.;
Berendsen, H. J. C. GROMACS: Fast, flexible, and free. J. Comput.
Chem. 2005, 26, 1701–1718.
(38) Valero, M. L.; Camarero, J. A.; Haack, T.; Mateu, M. G.; Domingo,
E.; Giralt, E.; Andreu, D. Native-like cyclic peptide models of a viral
antigenic site: finding a balance between rigidity and flexibility. J.
Mol. Recognit. 2000, 13, 5–13.
(39) van Regelmortel M. H. V.; Briand, J. P.; Muller, S.; Plaue, S. In
Laboratory techniques in biochemistry and molecular biology; Burdon,
R. H., Knippenberg, P. H., Eds.; Elsevier: Amsterdam, The Nether-
lands: 1988; pp 145–158.
(40) Buck, M. Trifluoroethanol and colleagues: cosolvents come of age.
Recent studies with peptides and proteins. Q. ReV. Biophys. 1998, 31,
297–355.
(41) Goodman, M.; Listowsky, I. Conformational aspects of synthetic
polypeptides VI. Hypochromic spectral studies of oligo-γ-methyl-L-
glutamate peptides. J. Am. Chem. Soc. 1962, 84, 3770–3771.
(57) Miyamoto, S.; Kollman, P. A. Settle - an analytical version of the
shake and rattle algorithm for rigid water models. J. Comput. Chem.
1992, 13, 952–962.
(58) ChemAxon MarvinBeans 4.1.0. http://www.chemaxon.com, 2007.
JM070196E