Secondary structure conversions of Alzheimer’s Ab(1–40)
peptide induced by membrane-mimicking detergents
Anna Wahlstro
¨m
1,
*, Loı
¨c Hugonin
1,
*, Alex Pera
´lvarez-Marı
´n
1,
*, Ju
¨ri Jarvet
2
and Astrid Gra
¨slund
1
1 Department of Biochemistry and Biophysics, The Arrhenius Laboratories for Natural Sciences, Stockholm University, Sweden
2 The National Institute of Chemical Physics and Biophysics, Tallinn, Estonia
Introduction
The amyloid bpeptide (Ab) is the major component of
the amyloid plaques, which are found in the brains of
Alzheimer’s disease patients. The Ab-peptide is a
39–42 residue peptide cleaved by processing of the
amyloid-bprecursor protein [1,2]. The Ab(1–40)
peptide has a hydrophilic N-terminal domain and a
more hydrophobic C-terminal domain, and contains a
central hydrophobic cluster (residues 17–21) suggested
to play an important role in peptide aggregation. Solu-
ble oligomeric peptide aggregates are reported to medi-
ate toxic effects on neurons and synapses [1,3] and
have attracted growing interest because of their proba-
ble link to the pathology of the disease. The formation
of aggregates occurs in parallel with a conformational
change of the peptide structure to b-sheet.
In vitro, the Abmonomer is in a dominating
random coil secondary structure in solution at room
temperature and physiological pH [4–7]. However, in
Keywords
amyloid bpeptide; CD; NMR; oligomer; SDS
Correspondence
A. Gra
¨slund, Department of Biochemistry
and Biophysics, The Arrhenius Laboratories
for Natural Sciences, Stockholm University,
SE-10691 Stockholm, Sweden
Fax: +46 8 155597
Tel: +46 8 162450
E-mail: astrid@dbb.su.se
*These authors contributed equally to this
work
(Received 29 April 2008, revised 8 August
2008, accepted 13 August 2008)
doi:10.1111/j.1742-4658.2008.06643.x
The amyloid bpeptide (Ab) with 39–42 residues is the major component of
amyloid plaques found in brains of Alzheimer’s disease patients, and solu-
ble oligomeric peptide aggregates mediate toxic effects on neurons. The Ab
aggregation involves a conformational change of the peptide structure to
b-sheet. In the present study, we report on the effect of detergents on the
structure transitions of Ab, to mimic the effects that biomembranes may
have. In vitro, monomeric Ab(1–40) in a dilute aqueous solution is weakly
structured. By gradually adding small amounts of sodium dodecyl sulfate
(SDS) or lithium dodecyl sulfate to a dilute aqueous solution, Ab(1–40) is
converted to b-sheet, as observed by CD at 3 C and 20 C. The transition
is mainly a two-state process, as revealed by approximately isodichroic
points in the titrations. Ab(1–40) loses almost all NMR signals at dodecyl
sulfate concentrations giving rise to the optimal b-sheet content (approxi-
mate detergent peptide ratio = 20). Under these conditions, thioflavin T
fluorescence measurements indicate a maximum of aggregated amyloid-like
structures. The loss of NMR signals suggests that these are also involved
in intermediate chemical exchange. Transverse relaxation optimized spec-
troscopy NMR spectra indicate that the C-terminal residues are more
dynamic than the others. By further addition of SDS or lithium dodecyl
sulfate reaching concentrations close to the critical micellar concentration,
CD, NMR and FTIR spectra show that the peptide rearranges to form a
micelle-bound structure with a-helical segments, similar to the secondary
structures formed when a high concentration of detergent is added directly
to the peptide solution.
Abbreviations
Ab-peptide, amyloid bpeptide; D P, detergent to peptide ratio; HSQC, heteronuclear single quantum coherence; LiDS, lithium dodecyl
sulfate; ppII, polyproline II; SDS-d
25,
deuterated SDS; ThT, thioflavin T; TROSY, transverse relaxation optimized spectroscopy.
FEBS Journal 275 (2008) 5117–5128 ª2008 The Authors Journal compilation ª2008 FEBS 5117
membrane-mimicking environments, such as SDS
micelles, the Ab-peptide displays an a-helical structure,
with two a-helical segments comprising residues 15–24
and 29–35, separated by a flexible hinge [8], and less
structured N- and C-termini. In the presence of phos-
pholipid vesicles, a-helical structures as well as b-sheet
structures have been reported [9]. Rangachari et al.
[10] have described interfacial aggregation of Ab(1–40)
at a polar nonpolar interface, with a concomitant
increase in b-structure content, brought about by SDS
micelles. In line with this finding, it was recently shown
that the Ab(1–40) and Ab(1–42) peptides form b-sheet-
rich aggregates at SDS concentrations significantly
below the critical micellar concentration [11]. These
aggregates give rise to thioflavin T (ThT) fluorescence
and are neurotoxic.
In the present study, we report on further properties
of soluble oligomeric b-sheet-rich Ab(1–40) aggregates,
formed at submicellar SDS or lithium dodecyl sulfate
(LiDS) concentrations at detergent peptide ratios of
approximately 20. The results obtained by CD, NMR,
FTIR and ThT fluorescence are compared and inter-
preted in terms of mixed micelle-like aggregates with
amyloid properties at intermediate detergent concen-
trations, where the peptides show dynamic properties,
particularly in the C-termini.
Results
Structural transitions of the Ab-peptide induced by
increasing concentrations of membrane-mimicking
detergents (SDS or LiDS) were studied by CD, NMR
and FTIR spectroscopy at temperatures in the range
3–25 C. LiDS was used at low temperature measure-
ments because it has a higher solubility at lower tem-
peratures than SDS; however, the critical micellar
concentration is approximately the same for the two
detergents [12,13].
CD spectroscopy
Detergent titration experiments were performed on a
sample with 75 lmAb(1–40) peptide in 10 mmsodium
phosphate buffer at 3 C and 20 C and pH 7.2. The
structural starting point for Ab(1–40) varies to some
extent as a function of temperature. At 3 C, the
secondary structure includes contributions from a poly-
proline II (ppII) helix, whereas, at 20 C, the second-
ary structure is almost exclusively random coil, as
described previously [5].
Figure 1A shows the titration of the Ab(1–40) pep-
tide with microvolumes of LiDS at 3 C over a deter-
gent concentration interval in the range 0.05–20 mm,
corresponding to detergent peptide (D P) ratios of
0.7–267, respectively. The CD data report on a first
structural conversion from a mixture of ppII helix and
random coil (weak positive shoulder at approximately
220 nm and negative minimum at 198 nm) occurring
at low LiDS concentrations (up to 0.7 mm,DP=9)
to a signal appearing at 1.6 mmLiDS (D P = 21)
with a maximum at 195 nm and a minimum at
218 nm, indicative of a dominating b-sheet structure.
It should be noted that, up to this titration point, the
spectra show a relatively well defined isodichroic point,
implying a two-state transition between the initial
structure and the b-sheet structure. After increasing
the LiDS concentration further (3.0 mmLiDS,
DP = 40), a new state is observed, mostly consisting
of a-helix structure. The conversion to a-helix struc-
ture reached its final state at 20 mmLiDS with a char-
acteristic maximum and two minima at 193 and
208 222 nm, respectively.
Figure 1B shows the SDS titration experiment at
20 C. In the absence of SDS, Ab(1–40) gives a CD
spectrum with a minimum at 198 nm, indicating a pre-
dominantly random coil secondary structure. As the
detergent concentration was increased, the CD signal
disappeared in the wavelength region around 198 nm
(SDS concentration of approximately 4 mm,
DP = 53). Further increase of the SDS concentration
(up to 5 mm,DP = 67) yielded a b-sheet spectrum
with a positive maximum at 194 nm and a negative
minimum at 218 nm. Also at this temperature, there
was a relatively well-defined isodichroic point in the
titration; however, this was not as clear as in the 3 C
titration. At high SDS concentrations (above 10 mm
SDS, D P = 133), the secondary structure was mainly
a-helix, with a characteristic maximum at 192 nm and
two minima at 208 and 221 nm.
The mean residual molar ellipticity at 195 nm as a
function of detergent concentration at 3 C and 20 C
is shown in Fig. 1C. The disappearance of an initial
weakly structured state and conversion to b-sheet and
then to a-helix are evident. The CD intensities at this
wavelength allowed us to compare the detergent
secondary structure induction at 3 C and 20 C
(Fig. 1C). Only one transition was visible with a mid-
point at 1 mmLiDS at 3 C. At 20 C and with SDS,
three transitions could be distinguished. The first had a
midpoint at 0.7 mm, followed by two more transitions,
with midpoints at 2.1 and 4.6 mmSDS.
Figure 1D shows the corresponding curves for the
mean residual ellipticity at 208 nm as function of
detergent concentration. At 3 C with LiDS, the data
show two sigmoidal transitions. The first sigmoidal
transition (positive) occurred in the range 0–1.6 mm
Detergent-induced Ab(1–40) secondary structures A. Wahlstro
¨met al.
5118 FEBS Journal 275 (2008) 5117–5128 ª2008 The Authors Journal compilation ª2008 FEBS
LiDS with a midpoint at 0.7 mm, corresponding to the
transition from initial structure to the structure domi-
nated by b-sheet. The second sigmoidal transition (neg-
ative) had a midpoint at 2.6 mm. We interpret this as
corresponding to the transition from b-sheet to a-helix.
In the SDS titration at 20 C, the intensity at 208 nm
again indicated three sigmoidal transitions. Two posi-
tive transitions had midpoints at 0.9 and 2 mm, respec-
tively, and the third (negative) had a midpoint at
6mm. It should be noted that three transitions are
visible at 20 C at both wavelengths studied, and that
the SDS concentration midpoints are in approximate
agreement: the first transition at approximately 0.8 mm
SDS (D P = 11), the second one at approximately
2mmSDS (D P = 27) and the third one at
approximately 5 mmSDS (D P = 67). The third
transition probably involves the formation of the
partly a-helical state, whereas the two first may involve
two similar but distinguishable states with b-sheet
structures.
Fig. 1. Circular dichroism spectra of 75 lMAb(1–40) peptide in 10 mMphosphate buffer at pH 7.2 in the presence of different concentra-
tions of detergent. (A) At 3 C in LiDS: open square, buffer; open circle, 0.05 mM; open triangle, 0.1 mM; filled square, 0.3 mM; open
diamond, 0.5 mM; filled circle, 0.7 mM; filled hexagon, 1.0 mM; open hexagon, 1.3 mM; open star, 1.6 mM; cross, 2.0 mM; filled star,
3.0 mM; open pentagon, 20 mM. (B) At 20 C in SDS: open square, buffer; open circle, 0.1 mM; filled star, 0.8 mM; open triangle, 2.0 mM;
open pentagon, 3.8 mM; filled square, 4.2 mM; open diamond, 4.3 mM; filled circle, 5.0 mM; filled triangle, 6.2 mM; open hexagon, 7.0 mM;
open star, 12.2 mM. (C) Plot of the mean residual molar ellipticity at 195 nm for the experiment in LiDS at 3 C (filled square) and for the
experiment in SDS at 20 C (open circle). (D) Plot of the mean residual molar ellipticity at 208 nm for the experiment in LiDS at 3 C (filled
square) and for the experiment in SDS at 20 C (open circle).
A. Wahlstro
¨met al. Detergent-induced Ab(1–40) secondary structures
FEBS Journal 275 (2008) 5117–5128 ª2008 The Authors Journal compilation ª2008 FEBS 5119
NMR spectroscopy
Heteronuclear single quantum coherence (HSQC) and
transverse relaxation optimized spectroscopy (TROSY)
NMR spectroscopy were used to follow the structural
transitions of the Ab(1–40) peptide (75 lm) induced by
increasing concentrations of the membrane-mimicking
detergent LiDS. The
1
H-
15
N HSQC spectrum of
uniformly
15
N-labeled Ab(1–40) in 10 mmphosphate
buffer (pH 7.2, 3 C) at the beginning of a titration is
shown in Fig. 2 (left). The corresponding spectrum of
the peptide in 128 mmLiDS at the end of a titration is
also shown in Fig. 2 (right, green spectrum). There are
significant chemical shift differences in comparison to
the initial state. Figure 2 (right) also includes the
HSQC spectrum of the peptide after direct addition of
150 mmLiDS at 3 C (red spectrum). The two spectra
shown in Fig. 2 (right) were found to overlap very well
with one another. However, the intensities (when
corrected for different peptide concentrations) were
significantly smaller in the spectrum after titration.
Assignments of the amide groups of Ab(1–40) in
buffer (Table S1) were made by comparison with the
previous assignment [14]. Assignment of Ab(1–40) in
150 mmLiDS at 3 C (Table S1) was performed by
starting the NMR experiment at 25 C where assign-
ments are known [8] and decreasing the temperature
by 5 C at a time following the gradual changes of the
HSQC spectra. The similarity of chemical shift
patterns at 3 C and 25 C suggests that the previously
determined a-helical regions involving residues 15–24
and 29–35 are the same at the two temperatures after
direct addition of a high concentration of detergent [8].
Between the two well defined states shown in Fig. 2
(i.e. at an intermediate detergent concentration), a new
state of the peptide characterized by complete NMR
signal loss was observed. This state occurred at a criti-
cal concentration of LiDS of 1–2 mm, corresponding
to D P = 13–27.
There was no obvious change in chemical shifts, nor
linewidth, of the amide HSQC crosspeaks by the grad-
ual titration with detergent below the concentration
inducing signal loss. To study how the signal was influ-
enced by an increasing concentration of detergent, the
volume of each crosspeak was integrated. In a titration
series with small titration steps (0.05, 0.1, 0.2, 0.3, 0.5,
0.7, 1, 2, 10 and 20 mm), most of the signals were
unchanged or slowly decayed up to a LiDS concentra-
tion of 0.5 mm. However, beyond 0.5 mmLiDS, the
signal from every residue abruptly decreased (Fig. 3).
Fig. 2. HSQC NMR spectra and assignment of amide crosspeaks for the Ab(1–40) peptide, and the effect of added lithium dodecyl sulfate.
Left:
1
H-
15
N HSQC spectrum of 75 lMuniformly
15
N-labeled Ab(1–40) in 10 mMphosphate buffer. The two peaks (V39 and V40) found in
the TROSY experiment with 75 lM
15
N-Ab(1–40) in the presence of 2 mMLiDS are indicated with arrows. Right: overlay of HSQC spectra;
75 lM
15
N-labeled Ab(1–40) in 128 mMof LiDS (i.e. the end point in the titration series 0, 0.5, 1, 4, 8, 16, 32, 64 and 128 mMLiDS) (green
spectrum) and 300 lM
15
N-labeled Ab(1–40) in 150 mMof LiDS, added in one addition (red spectrum). The peak intensities are corrected in
relation to the different peptide concentrations. All measurements were performed in 10 mMphosphate buffer at pH 7.2 and 3 C.
Detergent-induced Ab(1–40) secondary structures A. Wahlstro
¨met al.
5120 FEBS Journal 275 (2008) 5117–5128 ª2008 The Authors Journal compilation ª2008 FEBS
At 1 mmLiDS, corresponding to D P = 13, almost
all the HSQC crosspeaks had disappeared and, at
2mm, all were lost. At LiDS concentrations of 10 and
20 mm, the crosspeaks reappeared, directly with chemi-
cal shifts closely corresponding to those observed after
direct addition of 150 mmLiDS (Fig. 2, right, red
spectrum).
The crosspeaks from the amide groups in the amino
acid residues in the N- and C-terminal ends returned
with the strongest signals upon titration with detergent
(Fig. 3). This is probably due to an increased mobility
in the N- and C-terminal end segments (i.e. residues
up to G9 and beyond G37). The chemical shifts
observed at detergent concentrations of 10 and 20 mm
were retained in the presence of the higher LiDS con-
centrations of 64 and 128 mm, which all coincide with
the chemical shifts found at 150 mmLiDS (Fig. 2).
The disappearance of all NMR peaks at detergent
concentrations of 1–2 mmmay have more than one
explanation. An obvious reason for signal loss is that
Fig. 3. The crosspeak signal intensity of assigned residues of
15
N-labeled Ab(1–40) in the
1
H-
15
N HSQC spectra as a function of LiDS con-
centration (0, 0.05, 0.1, 0.2, 0.3, 0.5, 0.7, 1, 2, 10, 20 and 128 mM) at pH 7.2 and 3 C. The volumes of the HSQC crosspeaks of 75 lMof
15
N-Ab(1–40) were integrated. The amino acids are divided into different figures according to the earlier findings indicating that residues
15–24 and 29–35 have a-helical structure, whereas the regions in the ends and in between are unstructured [8]. The x-axis (LiDS concentra-
tion) is shown as a logarithmic scale.
A. Wahlstro
¨met al. Detergent-induced Ab(1–40) secondary structures
FEBS Journal 275 (2008) 5117–5128 ª2008 The Authors Journal compilation ª2008 FEBS 5121