Tài liệu Báo cáo khoa học: The Alzheimer b-peptide shows temperature-dependent transitions between left-handed 31-helix, b-strand and random coil secondary structures doc

The Alzheimer b-peptide shows temperature-dependent
transitions between left-handed 3
1
-helix, b-strand and
random coil secondary structures
Jens Danielsson, Ju
¨
ri Jarvet, Peter Damberg and Astrid Gra
¨
slund
Department of Biochemistry and Biophysics, Stockholm University, Sweden
The amyloid b-peptide (Ab) is the major component of
the amyloid plaques found in the extracellular com-
partment in the brains of patients suffering from
Alzheimer’s disease. The Ab-peptide is a 39–42-residue
peptide with the sequence: DAEFRHDSGYEVHHQ
KLVFFAEDVGSNKGAIIGLMVGGVVIA(1–42). It
is cleaved from the Alzheimer’s precursor protein by
the proteases b- and c-secretase [1,2]. The Ab(1–40)
peptide has a hydrophilic N-terminal region and a
more hydrophobic C-terminal region. The peptide con-
tains a central hydrophobic cluster, residues 17–21,
which is suggested to play an important role in peptide
aggregation [3]. There is experimental evidence that
soluble oligomeric aggregates have toxic effects on
neurons and synapses [1,4]. The aggregation involves a
conformational change of the peptide structure to
b-sheet. Solid state NMR spectroscopy has shown that
fibrils of Ab contain parallel b-sheet structure, whereas
shorter fragment fibrils consist of antiparallel b-sheet
structure [5,6]. In vitro, the Ab monomer is in a domi-

nating random coil secondary structure in solution at
room temperature and physiological pH [7–9].
Keywords
amyloid b-peptide; b-strand; left-handed
3
1
-helix; random coil; transition enthalpy
Correspondence
A. Gra
¨
slund, Department of Biochemistry
and Biophysics, Stockholm University,
S-106 91 Stockholm, Sweden
E-mail:
(Received 13 April 2005, revised 26 May
2005, accepted 9 June 2005)
doi:10.1111/j.1742-4658.2005.04812.x
The temperature-induced structural transitions of the full length Alzheimer
amyloid b-peptide [Ab(1–40) peptide] and fragments of it were studied
using CD and
1
H NMR spectroscopy. The full length peptide undergoes
an overall transition from a state with a prominent population of left-
handed 3
1
(polyproline II; PII)-helix at 0 °C to a random coil state at
60 °C, with an average DH of 6.8 ± 1.4 kJÆmol
)1
per residue, obtained by
fitting a Zimm–Bragg model to the CD data. The transition is noncoopera-

tive for the shortest N-terminal fragment Ab(1–9) and weakly cooperative
for Ab(1–40) and the longer fragments. By analysing the temperature-
dependent
3
J
HNHa
couplings and hydrodynamic radii obtained by NMR
for Ab(1–9) and Ab(12–28), we found that the structure transition includes
more than two states. The N-terminal hydrophilic Ab(1–9) populates PII-
like conformations at 0 °C, then when the temperature increases, confor-
mations with dihedral angles moving towards b-strand at 20 °C, and
approaches random coil at 60 °C. The residues in the central hydrophobic
(18–28) segment show varying behaviour, but there is a significant contri-
bution of b-strand-like conformations at all temperatures below 20 °C. The
C-terminal (29–40) segment was not studied by NMR, but from CD differ-
ence spectra we concluded that it is mainly in a random coil conformation
at all studied temperatures. These results on structural preferences and
transitions of the segments in the monomeric form of Ab may be related
to the processes leading to the aggregation and formation of fibrils in the
Alzheimer plaques.
Abbreviations
Ab-peptide, amyloid b-peptide; PII, polyproline II.
3938 FEBS Journal 272 (2005) 3938–3949 ª 2005 FEBS
An NMR study at 8 °C of the Ab(1–40) and Ab(1–
42) peptide with oxidized Met35 showed deviations
from random coil behaviour, but only limited informa-
tion about the solution structure could be derived [10].
The details of the high resolution structure or structure
propensity of Ab are still not well known. In order to
understand the early aggregation process and oligo-

merization of the peptide, further knowledge about the
structural energy landscape is needed, and this motiva-
ted the present study.
Earlier studies show that for the fragment Ab(12–
28) the secondary structure of the monomer changes
gradually towards a left-handed 3
1
(polyproline II;
PII)-helix when lowering the temperature [11]. Also the
fragment Ab(1–28) has been shown to adopt a PII-
helical structure in acidic solution [12]. For many short
peptides, however, not for all, the PII-helix seems to
dominate at low temperatures [13–15]. The PII-helix is
an extended structure with a rotational symmetry of
three amino acids per 360-degree turn. The torsion
angles are (/,w) ¼ ()78°, 146°). The PII-helix differs
from PI in the x-torsion angle, where PII is all trans
peptide bonds.
There are no interresidual hydrogen bonds stabil-
izing the PII structure. The stabilizing factor is pro-
posed to be interactions with the solvent. PII can exist
only in water and the structure is more stable in D
2
O
than in H
2
O, suggesting that water–peptide hydrogen
bonds are involved in the stabilization [14,16–18]. Dif-
ferent residues have different propensities to adopt the
PII-helix conformation. FTIR, Raman and different

CD experiments have led to estimations of the propen-
sity of the amino acids for the PII-helix [14]. Using
these results to predict the secondary structure of the
Ab-peptide shows that the PII content of the full
length peptide should be % 40% at low temperature
and predominantly in the N-terminal half of the pep-
tide [14]. The PII-helix is generally more stable at low
temperatures. The fraction of PII-helix increases as the
temperature decreases, an observation valid both for
true polyproline helices and other left-handed 3
1
-heli-
ces [11,13]. Raising the temperature induces a struc-
tural transition. This transition has been suggested to
be noncooperative for short peptides [19]. For longer
peptides molecular dynamics simulations of polyala-
nine suggest a cooperative transition [18]; however, the
theoretical results are dependent of the force field used
[20]. We have earlier observed that the Ab-peptide is
more soluble at low temperatures and is stable when
kept at low temperature [11,21] suggesting that PII-
helix prevents aggregation of the Ab-peptide.
The general properties of a PII-helix have been
determined by various spectroscopic methods such as
CD, NMR, FTIR and Raman optical activity [13]. In
CD spectroscopy a characteristic positive band appears
in the 210–230 nm region [13]. This positive band cor-
responds to an n–p* transition and is at 229 nm in
pure polyproline. It is shifted towards shorter wave-
lengths when other residues are involved [22]. The CD

spectra of PII-helices were earlier often interpreted as
random coil spectra. However, no positive bands or
local maxima should appear in a true random coil CD
spectrum [22]. The dihedral angles / and w that define
the PII structure are also reflected in the J couplings
between spins in the residues.
3
J
HNHa
couplings can be
studied by NMR spectroscopy. The
3
J
HNHa
coupling
between the amide proton and the a-carbon proton is
dependent on the torsion angle /.
The hydrodynamic properties of peptides reflect
structural properties. The hydrodynamic radius is rela-
ted to the diffusion coefficient via Stoke–Einstein’s
equation, and is dependent on not only the size but
also on the structure of the diffusing Ab-peptide and
scales with the molecular mass as the power law
R
H
¼ n
1
M
n2
[8]. The hydrodynamic radius is also rela-

ted to the persistence length, which depends on the
structural state of the peptide [23,24].
In the present investigation, we have explored the
structure propensities of Ab(1–40) and selected frag-
ment peptides. Using varying temperatures, the energy
landscape close to the solution structure is explored
and information on the structural transitions of the
peptide is obtained. The temperature-induced struc-
tural transitions also yield information on the back-
ground for a potential mechanism for the transition
from soluble monomer to aggregated multimer. We
have used CD as well as NMR at physiological pH, in
a temperature range from 0 °Cto60°C, to study the
solution structure of the peptide as well as the PII to
coil structural transition. The results show that the full
length peptide monomer partially adopts a PII-helix
structure at low temperatures, particularly in an N-ter-
minal region. However, the central hydrophobic clus-
ter, residues 17–21 and particularly the phenylalanine
residues 19 and 20, have a tendency towards b-strand
formation also at low temperature.
Results
Circular dichroism (CD) spectroscopy
Temperature dependence
We recorded CD spectra at varying temperatures for
the full length Ab(1–40)-peptide, as well as the frag-
ments Ab(1–9), Ab(1–16), Ab (1–28), Ab(12–28),
Ab(16–21), Ab(25–35) and the variant fragment
J. Danielsson et al. Structural transitions of Alzheimer b-peptide
FEBS Journal 272 (2005) 3938–3949 ª 2005 FEBS 3939

Ab(12–28)G19G20. For each peptide a CD spectrum
was obtained at a series of temperatures, ranging from
0to60°C. Figure 1 shows the temperature-dependent
CD spectra of all the studied peptides. Figure 1H
shows Ab(25–35) at 0 °C.
Among the fragments studied here, all N-terminal
peptides and the full length peptide are in monomeric
form at 10 lm concentrations, as shown earlier with dif-
fusion measurements [8]. To ensure that the temperature
dependence of the CD spectra reveals structural transi-
tions a control experiment was performed with 10 lm
Sml1(50–104) [25], a 54-residue unstructured peptide in
10 mm phosphate buffer at pH 7.2. This peptide shows
no structure transition. The CD spectrum shows only
small changes, as shown in Fig. S1, and these do not
follow the same pattern as the Alzheimer fragments.
From the CD experiments we conclude that the
N-terminal Alzheimer peptide fragments, as well as the
(1–40) peptide, are in equilibrium between a structural
state with a large contribution of PII-helix secondary
structure (most prominent at 0 °C) and random coil
structural state (most prominent at 60 °C). The two
state equilibrium is consistent with a relatively well-
defined isodichroic region around 208 nm.
Ab(25–35)
The hydrophobic fragment Ab(25–35) is different from
the other peptides in its CD characteristics. It is in an
aggregated b-sheet form under these conditions at all
temperatures between 0 °C and 60 °C, as the CD spec-
tra show little temperature variation (data not shown).

The shape of the spectrum recorded at 0 °C in Fig. 1H
suggests antiparallel b-sheet secondary structure. Mole-
cular mass studies confirmed the aggregated state of
the peptide as described below. This peptide was not
studied further.
PII content estimation
Quantification of the PII content of the CD spectra can
be carried out in several ways [26,27]. Here, the popula-
tion of PII-helix was calculated from the CD amplitude
of the local maximum around 220 nm, [h]
max
. Because
the wavelength of this maximum is dependent on the
sequence [13] we determined an individual [h]
max
for
each series of recorded spectra. From the spectral
intensities at [h]
max
the PII population, x
PII
was esti-
mated using the relation published by Kelly et al. [27]:
x
PII
¼
h½
max
þ 6100
13700

In the CD spectra of Ab(1–40) in Fig. 1A the local
maximum at 221 nm, best seen at low temperature, is
characteristic of a PII-helix. The population is strongly
temperature dependent and as the temperature reaches
60 °C only a small fraction of PII-helix remains. The
PII content was estimated to 43% at 0 °C and reduced
to 15% at 60 °C. At 37 °C the PII content was 20%.
The PII population at 0 °C is higher for the shorter
fragments compared to Ab(1–40) and ranges from
240220200
0
-10
velength / nm
[θ]/10
-3
deg cm
2
dmol
-1
240220200
0
-10
Wavelength / nm
240220200
0
-10
[θ]/10
-3
deg cm
2

dmol
-1
240220200
0
-10
-20
A
β
(1-40)
A
β
(1-28)
A
β
(1-16)
A
β
(12-28)
AB
CD
240220200
0
-10
[θ]/10
-3
deg cm
2
dmol
-1
240220200

40
20
0
-20
240220200
0
-10
[θ]/10
-3
deg cm
2
dmol
-1
240220200
0
-10
-20
E
F
G
H
A
β
(1-9)
A
β
(16-21)
A
β
(12-28)G

19
G
19
A
β
(25-35)
Fig. 1. Far UV CD spectra of the fragments in 10 mM sodium phos-
phate buffer at pH 7.4. The concentration was 10 l
M for all frag-
ments. Spectra were recorded at 0, 10, 20, 30, 40, 50 and 60 °C.
Generally the population of PII-helix decreased as the temperature
increased. (A) The full length peptide Ab(1–40); (B) Ab(1–28); (C)
Ab(1–16); (D) Ab(12–28); (E) Ab(1–9); (F) the central hydrophobic
cluster KLVFFA, Ab(16–21); (G) the variant fragment Ab(12–
28)G19G20. In (H) the C-terminal fragment Ab(25–35) fragment at
0 °C is shown, indicating antiparallel b-sheet secondary structure.
Structural transitions of Alzheimer b-peptide J. Danielsson et al.
3940 FEBS Journal 272 (2005) 3938–3949 ª 2005 FEBS
46% for the Ab(12–28) to almost 60% for the Ab(1–9)
fragment. The calculated populations are in good
agreement with the predicted populations based on
structural propensities of different residues [14]. The
structure transition is fully reversible for all peptides
(Fig. 1A–G). This was shown by lowering the tempera-
ture to 0 °C after the temperature increase and then
recording a new spectrum at 0 °C. This new spectrum
was identical to the original low temperature spectrum
(data not shown), indicating that no irreversible pro-
cesses are present.
We could estimate the location of the PII-helix on

the Ab(1–40) peptide by comparing the results on the
different fragments. The Ab(1–9) peptide shows high
propensity to form PII-helix, Fig. 1E. At 0 °C the pop-
ulation is 58%, and at high temperature it reduces to
42%. The central hydrophobic stretch, Ab(16–21)
shows a spectrum that has some distinct differences
compared to the other fragments, Fig. 1F. The evalu-
ated PII populations of the fragments except Ab(16–
21) are shown in Fig. 2.
As already mentioned, the Ab(23–35) fragment
aggregated under the conditions used here, and could
therefore not represent the monomeric form of the
C-terminal segment of Ab. However, considering the
differences observed between Ab(1–40) and Ab(1–28)
one can gain some information about the (29–40) seg-
ment of Ab. Supplementary Fig. S2 shows the differ-
ence spectra at 0 °C, 20 °C and 60 °C. The difference
spectra are dominated by a contribution from random
coil structure. At 0 °C there is a small contribution
(< 10%) from PII-helix.
Cooperativity for longer and noncooperativity
for shorter peptides
The amount of PII-helix, h
PII
, in the peptides is
strongly temperature dependent, and the temperature
dependence varies within the peptide group. From the
fitting of h
PII
according to the linearized model in

Eqns (5) and (6) to the data of Fig. 2, a transition
temperature T
m
, an enthaply change DH and a cooper-
ativity r was obtained for each peptide, Table 1. The
parameter T
m
is the temperature when 50% of the
secondary structure of the peptide is populated as
PII-helix.
The full length peptide and the longer fragments
exhibit a certain degree of cooperativity in the trans-
ition. The shorter peptides show a lower degree or no
cooperativity. The shortest fragment Ab(1–9) shows
no cooperativity at all (r ¼ 1.00) in agreement with
earlier observations that short segments do not exhibit
cooperativity in the transition from PII to random coil
[19].
The shorter N-terminal fragments have a higher T
m
,
i.e. are in a more stable PII conformation than the
full length peptide. The central hydrophobic stretch
Ab(16–21) shows a somewhat different pattern and the
temperature dependent population curve cannot be fit-
ted to Eqn (5). These results suggest that this peptide
deviates in its behaviour from the others. It should be
pointed out that these parameters were obtained using
a full Zimm–Bragg model. The results are not very
sensitive to absolute concentrations of the peptides. On

the other hand, using a simplified Zimm–Bragg model
(valid only for very long polymer chains) gives very
different values of the cooperativity coefficient r, and
is not applicable here.
CD spectroscopy is a low resolution structural
method. The CD results could be interpreted in terms
of a two state equilibrium between a more structured
state with large contributions from a PII-helix and a
random coil state. However, the small deviations from
isodichroic point behaviour, and the anomalous
400300200
1
0.8
0.6
0.4
0.2
0
Temperature / K
ω
PII
/ PII population
Fig. 2. Population of PII-helix as a function of temperature. The fit-
ted curves are calculated assuming the Zimm–Bragg model using a
linearized approximation for s close to the transition temperature.
The fragments measured are Ab(1–40) (h), Ab(1–28 (.), Ab(1–16)
(r), Ab(12–28) (s), Ab(12–28)G19G20 (d) and Ab(1–9) (e). The
shortest fragment shows no cooperativity in the transition, in con-
trast to the full length peptide.
J. Danielsson et al. Structural transitions of Alzheimer b-peptide
FEBS Journal 272 (2005) 3938–3949 ª 2005 FEBS 3941

behaviour of Ab(16–21) suggested that more structural
information might be obtained from NMR studies of
the peptides.
1
H NMR
J couplings and hydrodynamic radii
1
H NMR spectroscopy was used to study two shorter
fragments, Ab(1–9) and Ab(12–28) at varying tempera-
tures at 500 lm concentration and pH 7. Assignment
was based on standard procedures. The aim was to
obtain information on the temperature dependence of
the / angles along the peptide chain via
3
J
HNHa
cou-
plings. We also determined the temperature depend-
ence of the overall hydrodynamic radii for the
peptides. These experiments were carried out at higher
concentrations than those used for CD. However,
these shorter fragments are monomeric under the pre-
sent conditions, as verified by the diffusion experi-
ments. The diffusion coefficients were in good
agreement with what was expected from the molecular
masses of the monomers [8].
3
J
HNHa
couplings

The / dihedral angles differ for different secondary
structures: typical values are from )100° to )65° for
PII-helix and from )150° to )100° for b-strand. The
temperature dependences of the
3
J
HNHa
couplings
(data not shown) were studied for the two peptides.
The Karplus’ equation relates the / dihedral angle to
the
3
J
HNHa
coupling [28]. The / dihedral angle can be
determined from the inverse of Eqn (1) if / is between
)120° and )22°, according to
J
HNHa
¼ A þ B cosð/ À 60

ÞþC cos
2
ð/ À 60

Þð1Þ
Figure 3 shows the results for the / angles for selec-
ted residues obtained from
3
J

HNHa
couplings using
Bax’s parameters (A ¼ 1.60 Hz, B ¼ )1.76 Hz and
C ¼ 6.51 Hz) [29]. The shorter N-terminal fragment
Ab(1–9) shows a homogenous behaviour (Fig. 3A).
At low temperature all residues have angles corres-
ponding to a high population of PII-helix. When
raising the temperature the / angle moves towards
b-strand conformation (more negative values of /),
until a minimum of / is reached, after which the
values start to increase again. A random coil secon-
dary structure represents a weighted average of all
allowed dihedral angles and gives rise to
3
J
HNHa
couplings around 7 Hz. Formally this corresponds to
a single / angle of )80°. However, there are resi-
due-specific variations in the nature of the random
coil state [30]. These results suggest that NMR is
able to resolve an additional structural state for
certain residues in the Ab(1–9) peptide: besides the
PII-helix dominating at low temperature, there are
significant contributions from b-strand around 20–
30 °C, before random coil takes over around 50 °C.
Ala2 is an outlier, and has a much lower
3
J
HNHa
coupling than the other residues. This is in agree-

ment with other studies [19]. The small negative val-
ues of the evaluated / angles have been interpreted
as an indication of preferred PII conformation for
alanine.
The longer fragment Ab(12–28) shows a similar pat-
tern as Ab(1–9), but with some important differences
(Fig. 3B). Ala21 mainly follows the same pattern as
Ala2 in Ab(1–9). The other residues show a minimum
/ angle at 10–15 °C, after which they return towards a
random coil average. The two phenylalanines in the
central hydrophobic cluster, Phe19 and Phe20, have /
dihedral angles much shifted towards a b-strand con-
formation already at low temperatures. Phe19 and
Phe20 are believed to play an important role in peptide
aggregation. Figure 3 also includes the T
m
values
obtained by the two state interpretation of the CD
data. These values of T
m
appear to fall near the
minima of the / angle curves.
For comparison we also studied the variant peptide
Ab(12–28)G19G20. At low temperature, 277K, the
whole Ab(12–28)G19G20 peptide had larger
3
J
HNHa
couplings corresponding to more negative / angles
Table 1. Parameters calculated from the temperature dependence of the PII population. Cooperativity factor, given by r, where r ¼ 1isno

cooperativity and r << 1 is high cooperativity. Transition temperature (T
trans
) is in degrees Kelvin. DH is the enthalpy difference between PII-
helix and random coil (kJÆmol
)1
per residue). PII
pred
determined using the method proposed by Eker et al. [13] to predict the structure.
Ab(1–40) Ab(1–28) Ab(12–28) Ab(1–16) Ab(1–9) Ab(12–28)G19G20
r 0.14 0.32 0.26 0.40 1.00 0.21
T
trans
263 285 283 283 296 267
DH )4.0 )5.2 )6.3 )7.3 )8.7 )6.9
PII
max
0.45 0.54 0.46 0.52 0.57 0.44
PII
pred
0.40 0.57 0.47 0.62 0.67 0.52
Structural transitions of Alzheimer b-peptide J. Danielsson et al.
3942 FEBS Journal 272 (2005) 3938–3949 ª 2005 FEBS
(data not shown) than the native type peptide. Also
the shift towards random coil appeared at lower tem-
perature. These observations indicate that the trans-
ition towards b-strand occurs at lower temperatures in
the variant peptide than in Ab(12–28). This is also in
agreement with the observed values of T
m
from the

evaluations of the CD results (Table 1).
Hydrodynamic radii from translational diffusion
A structural transition changes the hydrodynamic
properties through the persistence length of the dif-
ferent structural groups. A true random coil has a
smaller hydrodynamic radius, R
H
, than a PII-helix
and PII has a smaller R
H
than an extended b-strand.
Figure 4 shows the calculated hydrodynamic radius
dependence on number of residues for a polypeptide
chain in ideal random coil, b-strand and PII-helix
conformations. Measurements of translational diffu-
sion for the Ab(12–28) and Ab(1–9) peptides were
performed at different temperatures. The use of ref-
erence molecules and the empirical function of the
viscosity as described below, gave the same results.
The hydrodynamic radii, R
H,
for the fragments
Ab(12–28) and Ab(1–9) as a function of temperature
are shown in Fig. 5. These results exhibit a similar
pattern of three structural states as the majority of
the
3
J
HNHa
couplings and support the interpretations

made in the previous section. When raising the tem-
perature, R
H
first increases reflecting the transition
from PII dominated to a b-strand containing state
with a more extended structure. At higher tempera-
tures R
H
decreases again due to higher populations
of random coil.
40200
40
30
20
10
0
N
o
of se
g
ments
R
H
/10
-10
m
Random Coil
β
-strand
Left-handed 3

1
-helix
Fig. 4. Simulated hydrodynamic radius for short peptide chains with
different secondary structures. For each chain length and each
structure 1500 structures were generated and the hydrodynamic
radius was calculated.
6040200
-60
-80
-100
φ-angle / degrees
3020100
A
B
T
m
A
2
H
6
E
3
D
7
S
8
R
5
F
4

A
21
L
17
E
22
V
24
F
19
F
20
V
18
T
m
Fig. 3. The temperature dependence of the
/ angle calculated from J
HNHa
couplings.
The transition temperature T
m
calculated
from CD data is indicated by an arrow. (A)
Residues 2–8 of Ab(1–9); (B) residues 17,
18, 19–22, 24 of Ab(12–28).
J. Danielsson et al. Structural transitions of Alzheimer b-peptide
FEBS Journal 272 (2005) 3938–3949 ª 2005 FEBS 3943
Discussion
The CD results for Ab(1–40) and the various frag-

ments presented in Fig. 1 indicate that we may consi-
der the full length peptide as composed of at least two
different segments with at least two different structural
states in a temperature-dependent equilibrium. At low
temperature (0 °C) the N-terminal half displays domin-
ant characteristics of a PII-helix, most prominently dis-
played by, e.g. the short N-terminal fragment Ab(1–9),
whereas the C-terminal half is dominated by random
coil already at low temperature seen directly by com-
paring the CD spectra of Ab(1–40) and Ab(1–28),
Fig. 1A,B. At high temperature (60 °C) the whole pep-
tide is in a random coil state. The CD results presented
here are in agreement with a two-segment model, e.g.
with 65% PII propensity in the Ab(1–25) segment and
0% PII propensity in the Ab(26–40) segment at 0 °C.
The anomalous behaviour of Ab(25–35) being in an
aggregated b-state already at 0 °C indicates that the
less hydrophobic N-terminus helps to keep this part in
a monomeric state in the full length peptide.
The enthalpy change of the transition from PII to
random coil is calculated to 6.8 ± 1.4 kJÆmol
)1
per
residue, an average for all studied peptides. This
enthalpy change is less than that of a typical hydrogen
bond. This is in agreement with the lack of hydrogen
bonding to stabilize a PII conformation. Instead, a
PII-helix is considered to be stabilized by solvation in
an aqueous solution [18,31]. A conformational energy
map calculated for a hydrated alanine residue shows a

rather deep minimum for a region that encompasses
both b-sheet and PII-helix [32–34]. This suggests that
there is a low barrier between a generic b-sheet struc-
ture [35] and PII. Futhermore, the low energy barrier
may provide an explanation for a direct conversion
from one form to the other, concomitant with peptide
aggregation, if the concentration is high enough.
The cooperativity coefficient r of Ab(1–40) was
found to be 0.14, a relatively weak cooperativity com-
pared to, e.g. the a-helix forming 1500 residue
poly(benzyl-l-glutamate), for which a r of 2 · 10
)4
was reported [36]. The origin of the weak cooperativity
in Ab(1–40) may lie in interactions between neighbour-
ing large hydrophobic side chains.
The NMR results refine the segmental model of
Ab(1–40) and suggest three distinguishable segments of
the peptide, and three structural states in the tempera-
ture dependent equilibrium. All structural states have
large contribution of random coil, but some structural
preferences are indicated by the varying / angles for
the residues. A smaller negative / indicates PII-helix
and a larger negative / indicates b-strand conforma-
tion. The studies of the
3
J
HNHa
couplings in the two
fragments Ab(1–9) and Ab(12–28) (Fig. 3) indicate
that residues 2–8 begin in a PII-rich average conforma-

tion at 0 °C, move towards a b-strand conformation at
20–30 °C, and finally towards random coil at 60 °C.
Residues 17–24 display a more complex temperature
behaviour and do not all behave similarly. Two of
them (Leu17, Glu22) follow a pattern of PII fi
b-strand fi coil similar to residues 2–8, whereas Ala21
goes directly from PII-rich state to random coil. Val18,
Phe19, Phe20 and Val24 on the other hand seem to
start in a b-strand-rich conformation already at 0 °C,
and go directly from there to random coil. This high
b-strand propensity is in good agreement with the pre-
dicted structures [14]. The temperature dependence of
the hydrodynamic radius R
H
supports the existence of
403020100
12
10
8
6
4
Temperature /
o
C
R
H
/10
-10
m
A

β
(1-9)
A
β
(12-28)
α
-CD
Fig. 5. The hydrodynamic radius as a function of temperature for
the fragment Ab(1–9) (s) and Ab(12–28) (d). The hydrodynamic
radius is calculated from pulsed field gradient NMR diffusion data
obtained at 500 l
M peptide concentration in 10 mM sodium phos-
phate buffer at pH 7. The temperature dependence of the hydrody-
namic radius is similar to that of the J couplings. The hydrodynamic
radius is calculated from the diffusion coefficient via Stoke–
Einstein’s equation. All data is corrected for temperature induced
viscosity changes. The temperature dependence of the hydro-
dynamic radius of a-cyclodextrin is shown as a reference (h).
Structural transitions of Alzheimer b-peptide J. Danielsson et al.
3944 FEBS Journal 272 (2005) 3938–3949 ª 2005 FEBS
three structural states of the peptide considered as a
whole.
We have no NMR results directly describing the
structural state of the C-terminal (29–40) segment.
However, the CD difference spectra between Ab(1–40)
and Ab(1–28) at different temperatures suggest that
the (29–40) segment is dominated by random coil
structure at all studied temperatures. There is a small
contribution of PII at 0 °C. Trying to break the struc-
tural states down into three different peptide segments,

(1–17), (18–24), (25–40), we arrive at a simplified
model with overall characteristics as described in
Fig. 6.
This structural state model has several interesting
features, in agreement with observed experimental
results. The PII contribution in the N-terminus helps
to keep the peptide in solution at low temperatures, in
agreement with our previous observations [21] of the
increased solubility of the peptide at low temperatures.
The gradual transition to more b-strand in the whole
N-terminus when the temperature is raised increases
the total content of b-strand in the sequence. If the
concentration is high enough this could trigger the
irreversible aggregation concomitant with the conver-
sion into b-structure, the favoured secondary structure
of the aggregates. The central segment 18–24, which
has a large contribution of b-strand already at the low
temperatures may be considered as the seed in this
structural transition of the peptide. It is already known
that the Ab(1–40)G19G20 variant does not aggregate
like the native sequence [37]. From the present results
we may suggest that Val18, Phe19, Phe20 and Val24
contribute considerably to the structural transition
towards b-structure, whereas Ala21 and Glu22 coun-
teract this transition by occupying a more tempera-
ture-stable PII conformation. Some known mutations,
Flemish (A21G), Dutch (E22Q), Italian (E22K) and
Arctic (E22G) associated with early onset of Alzhei-
mer’s disease are located at these residues [38].
This study is an attempt to combine various spectro-

scopic observations to give a realistic description of
what may in a very approximate description be called
an unstructured (random coil) peptide. Earlier studies
on the Ab-peptide structure in aqueous solution by
NMR have made use of observations at a single tem-
perature of
1
H
1
H NOEs,
3
J
HNHa
couplings and
15
N
1
H
NOEs. The results were presented as a model described
as a collapsed coil [39] or as a structure deviating from
random coil behaviour by local conformational prefer-
ences of short segments [10]. The present results have
been obtained avoiding NOE observations that may
lead to biased results in a highly flexible system. We
have made combined use of the temperature depend-
ence of the observed CD and NMR parameters for the
full length peptide as well as selected fragments, and
have taken a first step towards characterization with
atomic resolution of the structure transitions that
occur in Ab(1–40). The observations may be helpful

for understanding the Alzheimer peptide early aggrega-
tion behaviour.
Experimental procedures
The peptides, the full length Ab(1–40) and the fragments
Ab(1–28), Ab(1–16), Ab(12–28), Ab(1–28)G19G20, Ab(25–
35) and Ab(1–9) were purchased from Neosystem (Stras-
bourg, France). The samples were purified by HPLC by the
supplier and were used without further purification. All pep-
tides were nonmodified in the termini. The peptides were
stored at )18 °C and thawed before use. The CD experi-
ments were performed in 10 mm sodium phosphate buffer
at pH 7.4 at 10 lm peptide concentrations. The pH was
adjusted using phosphate buffer and was measured using a
standard pH-meter. The sample preparation was carried out
at 5 °C and the samples were kept at low temperature until
the experiments were performed. The concentrations of the
samples were determined by weight.
Circular dichroism
CD spectra were obtained with a Jasco (Easton, MO, USA)
J-720 spectropolarimeter and the temperature was controlled
with a PTC-343 temperature controller. A quartz cell with
2 mm optical path was used. The spectral range was 190–
250 nm with a resolution of 0.2 nm and a bandwidth of 2 nm.
A scan speed of 50 nmÆmin
)1
with 2 s response time was
employed. The background spectrum was subtracted and the
results were expressed as mean residue molar ellipticity [h].
Nuclear magnetic resonance
NMR experiments were performed on a Bruker Avance

TM
(Karlsruhe, Germany) 400 MHz spectrometer, a Bruker
Avance
TM
500 MHz spectrometer equipped with a cryo-
probe, Varian (Palo Alto, CA, USA) 600 and 800 MHz
DAEFR
5
HDSGY
10
EVHHQ
15
KLVFF
20
AEDVG
25
SNKGA
30
IIGLM
35
VGGVV
40
DAEFR
5
HDSGY
10
EVHHQ
15
KLVFF
20

AEDVG
25
SNKGA
30
IIGLM
35
VGGVV
40
DAEFR
5
HDSGY
10
EVHHQ
15
KLVFF
20
AEDVG
25
SNKGA
30
IIGLM
35
VGGVV
40
0°C
60°C
15-20°C
PII
β-strand
β-strand

coil
coil
coil
Fig. 6. A model of three distinct segments of Ab(1–40) as a mono-
mer in aqueous solution, and their dominating structure states at
different temperatures.
J. Danielsson et al. Structural transitions of Alzheimer b-peptide
FEBS Journal 272 (2005) 3938–3949 ª 2005 FEBS 3945
spectrometers. The sample concentrations were 500 l m for
both peptides, Ab(1–9) and Ab(12–28), in 10 mm sodium
phosphate buffer at pH 7.4. The temperature was calibrated
using a standard sample of 100% ethylene glycol. All
experiments were performed with samples in 90% H
2
O and
10% (v ⁄ v) D
2
O or 100% D
2
O (for diffusion experiments).
Diffusion experiments were performed using a pulsed field
gradient sequence with longitudinal storage and eddy cur-
rent delay, PFGLED. Solvent suppression was managed
with presaturation of the water. Diffusion experiments were
performed using a list of 32 gradient strengths, a gradient
pulse-length of 4 ms and a diffusion time of 100 ms. To
account for a nonlinear gradient profile the method des-
cribed by Damberg et al. was used [40].
3
J

HNHa
couplings were measured using standard 1D
1
H
NMR spectra when possible. Due to overlap 2D
1
H COSY
experiments were necessary when measuring
3
J
HNHa
cou-
plings on the longer fragment Ab(12–28). To obtain the
3
J
HNHa
couplings from the spectra a Lorentzian function
was fitted to the in-phase doublet peaks from the 1D
experiment and the anti-phase doublet peak from the 2D
1
H COSY. Water suppression in the 1D spectra was
obtained by using a jump-return sequence to avoid signal
loss due to water–amide proton exchange. In the 2D
1
H
COSY experiments a watergate sequence was used.
Viscosity correction in diffusion experiments
The diffusion coefficient is temperature dependent, primar-
ily because the thermal Brownian motion is directly propor-
tional to the temperature but also because the viscosity of

the solvent, here water, is highly temperature dependent,
which should be corrected for. To account for the viscosity
effect two different approaches were taken. First, reference
molecules with a known hydrodynamic radius were studied.
Here HDO and a-cyclodextrin were used as references. The
unknown hydrodynamic radius of the peptide was calcula-
ted using:
R
H
(T) ¼
D
ref
(T)
D
ref
ð298KÞ
Á
D
OBS
ð298K)
D
OBS
(T)
R
H;ref
D
ref
(298K) is the diffusion coefficient of the reference mole-
cule at a reference temperature T ¼ 298K, D
OBS

(298K) is the
measured peptides diffusion at that temperature. R
H,ref
is the
reference molecule’s hydrodynamic radius at the reference
temperature. D
ref
(T) and D
OBS
(T) are the diffusion coeffi-
cients at temperature T for the reference molecule and the
studied peptide, respectively. In a second approach, empirical
functions for the viscosity of H
2
O and D
2
O were obtained
from fitting tabulated values to the empirical function [41]:
g ¼ Ke
ÀjðTÀ273Þ
m
This approach gave the parameters (L, j, t) ¼ (2.41 · 10
)3
,
0.057, 0.8078) for D
2
O and (L, j, t) ¼ (1.80 · 10
)3
, 0.049,
0.8222) for H

2
O. Different mixtures of D
2
O and H
2
O
were then considered to have the weighted mean of the
viscosities [8].
Peptide aggregation
Aggregation was determined by molecular mass filtering
experiments. Light absorption at 212 nm was measured for
the fragments Ab(1–16) and Ab(25–35). The CD sample
solutions were centrifuged (2500 g,4°C, 20 min) through a
10 kDa cut-off filter and the absorption was measured
again. For the fragment Ab(1–16) no loss of peptide was
observed, but for the Ab(25–35) a loss of > 90%, indica-
ting severe aggregation of the peptide. Repeating this pro-
cedure but at low pH where the peptide is less prone to
aggregate shows significantly less loss of peptide, < 30%.
Structural transition theory
The temperature induced transition between PII-helix and
random coil can be treated as an ordinary helix–coil trans-
ition described by Zimm and Bragg. Here we use a slightly
modified Zimm–Bragg model [36]. Using their matrix
method, the partition function, Q, of a peptide with N resi-
dues is given by:
Q ¼ bM
NÀ1
a
0

ð2Þ
Here, a
0
is the vector describing only one residue, which is
assumed to be unstructured, b is a row vector consisting
only of ones. M is the matrix operator that adds one resi-
due to the vector of possible conformations with their sta-
tistical mass. For a system where M is a 2 · 2 matrix, as it
is here, Eqn (2) becomes:
Q ¼ 11½
1 rs
1 s

NÀ1
1
0

ð3Þ
s is the intrinsic statistical weight for a residue to be in PII
conformation. r is a cooperativity coefficient. The factor
rs is then the statistical weight for a residue PII conforma-
tion following after an unstructured residue. If r is 1 there
is no cooperativity in the transition, and if r ( 1 the coop-
erativity is high. Calculating the partition function from
Eqn (3) gives:
Q ¼
k
N
0
ðk

0
À sÞþk
N
1
ðs À k
1
Þ
k
0
À k
1
k
0;1
¼
1
2
1 þ s Æ
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ð1 þ sÞ
2
þ 4rs
q

ð4Þ
Here the eigenvalues k
0,1
of M are introduced. At equili-
brium the fraction of residues in PII conformation is given
by the power of s in the partition function divided by the
total number of residues. The fraction of PII is then x

PII
¼
(N-1)
)1
d(lnQ) ⁄ d(lns) and can be calculated from Eqn (4).
Performing this calculation we obtain:
Structural transitions of Alzheimer b-peptide J. Danielsson et al.
3946 FEBS Journal 272 (2005) 3938–3949 ª 2005 FEBS
x
PII
¼
s
Nc
0
k
0
þ
c
0
À1
k
0
Às

k
N
0
ðk
0
À sÞþ

Nc
1
k
1
þ
c
1
À1
k
1
Às

k
N
1
ðs À k
0
Þ

ðN À1Þðk
N
0
ðk
0
À sÞþk
N
1
ðs À k
1
ÞÞ

À
sðc
0
À c
1
Þ
ðN À1Þðk
0
À k
1
Þ
ð5Þ
c
0,1
are the derivatives of k
0,1
with respect to s:
c
0;1
¼
1
2
1 À
1 þ s þ 2r
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ð1 þ sÞ
2
þ 4rs
q
8

>
<
>
:
9
>
=
>
;
ð6Þ
The parameter s is, as mentioned above, related to the equi-
librium constant and thus to the enthalpy change due to
the structural transition.
The temperature dependence of the transition can be
studied, and if the populations of the structural entities can
be determined, the parameters s and r can be determined
from Eqn (5). Close to the transition temperature T
m
the
parameter s may be approximated as a linear function of
T [36].
s ¼ 1 À
DH
RT
2
m
ðT À T
m
Þ
At s ¼ 1 the two states are equally populated.

Hydrodynamic radius from simulations
Simulations were performed to obtain hydrodynamic data. As
a model for the peptides a peptide chain with no side chains
was used. Three bonds with fixed bond lengths represented
every residue. The bond lengths used were 1.33, 1.45 and
1.52 A
˚
and the fixed angles between the bonds were: 58°,64°
and 75°. The rotations around the bonds were controlled: for
a random coil peptide the rotation was free and for defined
structures the rotation was constrained. The simulation was
performed using a hard sphere model with a 1.3 A
˚
van der
Waals radius of the heavy atoms. Fifteen hundred structures
of each length were generated with different distributions of
torsion angles. The random coil had free rotation in all three
torsion angles. The PII-helix was simulated using equally dis-
tributed angles centred at (/,w,x) ¼ ()78°,146°, 180°) and
with a arbitrary width of 10° for all angles. A b-strand was
simulated in the same way with the torsion angles centred at
(/,w,x) ¼ ()139°,146°, 180°) and with the same flexibility.
The hydrodynamic radius was calculated as the radius of
gyration, the mean distance from the centre of mass.
Acknowledgements
This study was supported by a grant from the Swedish
Research Council and by the European Commission,
contracts LSHG-CT-2004-51 and QLK3-CT-2002-
01989. We wish to thank Maria Yamout for giving us
the Sml1 peptide sample.

References
1 Hardy J & Selkoe DJ (2002) The amyloid hypothesis of
Alzheimer’s disease: progress and problems on the road
to therapeutics. Science 297, 353–356.
2 Kim JE & Lee M (2003) Fullerene inhibits b-amyloid
peptide aggregation. Biochem Biophys Res Commun 303,
576–579.
3 Balbach JJ, Ishii Y, Antzutkin ON, Leapman RD, Rizzo
NW, Dyda F, Reed J & Tycko R (2000) Amyloid fibril
formation by Ab 16–22, a seven-residue fragment of the
Alzheimer’s b-amyloid peptide, and structural characteri-
zation by solid state NMR. Biochemistry 39, 13748–
13759.
4 Kuo YM, Webster S, Emmerling MR, De Lima N &
Roher AE (1998) Irreversible dimerization ⁄ tetrameriza-
tion and post-translational modifications inhibit proteo-
lytic degradation of Ab peptides of Alzheimer’s disease.
Biochim Biophys Acta 1406, 291–298.
5 Antzutkin ON, Leapman RD, Balbach JJ & Tycko R
(2002) Supramolecular structural constraints on Alzhei-
mer’s b-amyloid fibrils from electron microscopy and
solid-state nuclear magnetic resonance. Biochemistry 41,
15436–15450.
6 Petkova AT, Buntkowsky G, Dyda F, Leapman RD,
Yau WM & Tycko R (2004) Solid state NMR reveals
a pH-dependent antiparallel b-sheet registry in fibrils
formed by a b-amyloid peptide. J Mol Biol 335, 247–
260.
7 Ma K, Clancy E, Zhang Y, Ray D, Wollenberg K &
Zagorski M (1999) Residue-specific pK

a
measurements
of the b-peptide and mechanism of pH-induced amyloid
formation. J Am Chem Soc 121, 8698–8706.
8 Danielsson J, Jarvet J, Damberg P & Gra
¨
slund A
(2002) Translational diffusion measured by PFG-NMR
on full length and fragments of the Alzheimer Ab(1–40)
peptide. Determination of hydrodynamic radii of ran-
dom coil peptides of varying length. Magnetic Reson-
ance Chem 40, S89–S97.
9 Hou L, Shao H, Zhang Y, Li H, Menon NK, Neuhaus
EB, Brewer JM, Byeon IJ, Ray DG, Vitek MP et al.
(2004) Solution NMR studies of the Ab(1–40) and
Ab(1–42) peptides establish that the Met35 oxidation
state affects the mechanism of amyloid formation. JAm
Chem Soc 126, 1992–2005.
10 Riek R, Gu
¨
ntert P, Do
¨
beli H, Wipf B & Wu
¨
thrich K
(2001) NMR studies in aqueous solution fail to identify
significant conformational differences between the
monomeric forms of two Alzheimer peptides with
widely different plaque-competence, Ab(1–40) (ox) and
Ab(1–42) (ox). Eur J Biochem 268, 5930–5936.

11 Jarvet J, Damberg P, Danielsson J, Johansson I, Eriks-
son LE & Gra
¨
slund A (2003) A left-handed 3
1
helical
conformation in the Alzheimer Ab(12–28) peptide.
FEBS Lett 555, 371–374.
J. Danielsson et al. Structural transitions of Alzheimer b-peptide
FEBS Journal 272 (2005) 3938–3949 ª 2005 FEBS 3947
12 Eker F, Griebenow K & Schweitzer-Stenner R (2004)
Ab (1–28) fragment of the amyloid peptide predomin-
antly adopts a polyproline II conformation in an acidic
solution. Biochemistry 43, 6893–6898.
13 Shi Z, Woody RW & Kallenbach NR (2002) Is polypro-
line II a major backbone conformation in unfolded
proteins? Adv Protein Chem 62, 163–240.
14 Eker F, Griebenow K, Cao X, Nafie LA & Schweitzer-
Stenner R (2004) Preferred peptide backbone confor-
mations in the unfolded state revealed by the
structure analysis of alanine-based (AXA) tripeptides in
aqueous solution. Proc Natl Acad Sci USA 101,
10054–10059.
15 Blanch EW, Morozova-Roche LA, Cochran DA, Doig
AJ, Hecht L & Barron LD (2000) Is polyproline II helix
the killer conformation? A Raman optical activity study
of the amyloidogenic prefibrillar intermediate of human
lysozyme. J Mol Biol 301, 553–563.
16 Whittington SJ & Creamer TP (2003) Salt bridges do
not stabilize polyproline II helices. Biochemistry 42,

14690–14695.
17 Chellgren BW & Creamer TP (2004) Effects of H
2
O
and D
2
O on polyproline II helical structure. J Am Chem
Soc 126, 14734–14735.
18 Garcia A (2004) Characterization of non-alpha helical
conformations in Ala peptides. Polymer 45, 669–676.
19 Chen K, Liu Z & Kallenbach NR (2004) The poly-
proline II conformation in short alanine peptides is
noncooperative. Proc Natl Acad Sci USA 101, 15352–
15357.
20 Gnanakaran S & Garcı
´
a AE (2005) Helix-coil transition
of alanine peptides in water: force field dependence on
the folded and unfolded structures. Proteins 59,
773–782.
21 Jarvet J, Damberg P, Bodell K, Eriksson G & Gra
¨
slund
A (2000) Reversible random coil to b-sheet transition
and the early stage of aggregation of the Ab(12–28)
fragment from the alzheimer peptide. J Am Chem Soc
122, 4261–4268.
22 Chellgren BW & Creamer TP (2004) Short sequences of
non-proline residues can adopt the polyproline II helical
conformation. Biochemistry 43, 5864–5869.

23 Syme CD, Blanch EW, Holt C, Jakes R, Goedert M,
Hecht L & Barron LD (2002) A Raman optical
activity study of rheomorphism in caseins, synucleins
and tau. New insight into the structure and behaviour
of natively unfolded proteins. Eur J Biochem 269 ,
148–156.
24 Krieger F, Fierz B, Bieri O, Drewello M & Kiefhaber T
(2003) Dynamics of unfolded polypeptide chains as
model for the earliest steps in protein folding. J Mol
Biol 332, 265–274.
25 Zhao X, Muller EG & Rothstein R (1998) A suppressor
of two essential checkpoint genes identifies a novel
protein that negatively affects dNTP pools. Mol Cell 2,
329–340.
26 Bienkiewicz EA, Moon Woody A & Woody RW (2000)
Conformation of the RNA polymerase II C-terminal
domain: circular dichroism of long and short fragments.
J Mol Biol 297, 119–133.
27 Kelly MA, Chellgren BW, Rucker AL, Troutman JM,
Fried MG, Miller AF & Creamer TP (2001) Host-guest
study of left-handed polyproline II helix formation.
Biochemistry 40, 14376–14383.
28 Cavanagh J, Fairbrother W, Palmer A III & Skelton N
(1996) Protein NMR Spectroscopy, Principle and Prac-
tice. Academic Press, London.
29 Vuister GW & Bax A (1993) Quatitative J correlation: a
new approach for measuring homonuclear three-bond J
(HNHa) coupling constants in
15
N-enriched proteins.

J Am Chem Soc 115, 7772–7777.
30 Bundi A & Wu
¨
thrich K (1979)
1
H-NMR parameters of
the common amino acid residues measured in aqueous
solutions of the linear tetrapeptides H-Gly-Gly-X-L-
Ala-OH. Biopolymers 18, 285–297.
31 Woody RW (1992) Circular dichroism and conforma-
tion of unordered polypeptides. Adv Biophys Chem 2,
37–79.
32 Brant DA, Miller WG & Flory PJ (1967) Conforma-
tional energy estimates for statistically coiling polypep-
tides. J Mol Biol 23, 47–65.
33 Gnanakaran S & Garcia AE (2003) Validation of an
all-atom protein force field: from dipeptides to larger
peptides. J Phys Chem 107, 12555–12557.
34 Han WG, Jalkanen KJ, Elstner M & Suhai S (1998)
Theoretical study of aqueous N-acetyl-l-alanine
N¢-methylamide: Structures and Raman, VCD, and
ROA spectra. J Phys Chem B 102, 2587–2602.
35 Chamberlain AK, MacPhee CE, Zurdo J, Morozova-
Roche LA, Hill HA, Dobson CM & Davis JJ (2000)
Ultrastructural organization of amyloid fibrils by atomic
force microscopy. Biophys J 79, 3282–3293.
36 Zimm BH & Bragg JK (1959) Theory of the phase tran-
sition between helix and random coil in polypeptide
chains. J Chem Physics 31, 526–535.
37 Mansfield SL, Jayawickrama DA, Timmons JS &

Larive CK (1998) Measurement of peptide aggre-
gation with pulsed-field gradient nuclear magnetic
resonance spectroscopy. Biochim Biophys Acta 1382,
257–265.
38 Pa
¨
ivio
¨
A, Jarvet J, Gra
¨
slund A, Lannfelt L & Westlind-
Danielsson A (2004) Unique physicochemical profile of
b-amyloid peptide variant Ab1–40E22G protofibrils:
conceivable neuropathogen in arctic mutant carriers.
J Mol Biol 339, 145–159.
39 Zhang S, Iwata K, Lachenmann MJ, Peng JW, Li S,
Stimson ER, Lu Y, Felix AM, Maggio JE, Lee JP
(2000) The Alzheimer’s peptide aaˆ adopts a col-
Structural transitions of Alzheimer b-peptide J. Danielsson et al.
3948 FEBS Journal 272 (2005) 3938–3949 ª 2005 FEBS
lapsed coil structure in water. J Struct Biol 130,
130–141.
40 Damberg P, Jarvet J & Gra
¨
slund A (2001) Accurate
measurement of translational diffusion coefficients: a
practical method to account for nonlinear gradients.
J Magn Reson 148, 343–348.
41 Cho C, Urquidi J, Singh S & Robinson G (1999) Ther-
mal offset viscosities of liquid H

2
O, D
2
O and T
2
O.
J Phys Chem 103, 1991–1994.
Supplementary material
The following supplementary material is available
online:
Fig. S1. Far UV CD spectra of 10 lM Sml1 in
10 mM sodium phosphate buffer at pH 7.4.
Fig. S2. Far UV CD difference spectra between 10
lMAb(1–40) and Ab(1–28) in sodium phosphate buf-
fer at pH 7.4.
J. Danielsson et al. Structural transitions of Alzheimer b-peptide
FEBS Journal 272 (2005) 3938–3949 ª 2005 FEBS 3949