Modulation of fibrillation of hIAPP core fragments by chemical modification of the peptide backbone

Article abstract of DOI:10.1016/j.bbapap.2011.10.014

The well-ordered cross β-strand structure found in amyloid aggregates is stabilized by many different side chain interactions, including hydrophobic interactions, electrostatic charge and the intrinsic propensity to form β-sheet structures. In addition to the side chains, backbone interactions are important because of the regular hydrogen-bonding pattern. β-Sheet breaking peptide analogs, such as those formed by N-methylation, interfere with the repetitive hydrogen bonding pattern of peptide strands. Here we test backbone contributions to fibril stability using analogs of the 6-10 residue fibril core of human islet amyloid polypeptide, a 37 amino acid peptide involved in the pathogenesis of type II diabetes. The Phe-Gly peptide bond has been replaced by a hydroxyethylene or a ketomethylene group and the nitrogen-atom has been methylated. In addition, we have prepared peptoids where the side chain is transferred to the nitrogen atom. The backbone turns out to be extremely sensitive to substitution, since only the minimally perturbed ketomethylene analog (where only one of the - NH - groups has been replaced by - CH ₂-) can elongate wildtype fibrils but cannot fibrillate on its own. The resulting fibrils displayed differences in both secondary structure and overall morphology. No analog could inhibit the fibrillation of the parent peptide, suggesting an inability to bind to existing fibril surfaces. In contrast, side chain mutations that left the backbone intact but increased backbone flexibility or removed stabilizing side-chain interactions had very small effect on fibrillation kinetics. We conclude that fibrillation is very sensitive to even small modifications of the peptide backbone.

Full text of DOI:10.1016/j.bbapap.2011.10.014

Biochimica et Biophysica Acta 1824 (2012) 274–285
Contents lists available at SciVerse ScienceDirect
Biochimica et Biophysica Acta
journal homepage: www.elsevier.com/locate/bbapap
Modulation of fibrillation of hIAPP core fragments by chemical modification of the
peptide backbone
Maria Andreasen a, Søren B. Nielsen a, Tina Mittag b, Morten Bjerring b, Jakob T. Nielsen b, Shuai Zhang c,
Erik H. Nielsen b, Martin Jeppesen a, Gunna Christiansen d, Flemming Besenbacher c, Mingdong Dong c,
a,
Niels Chr. Nielsen b, Troels Skrydstrup b, , Daniel E. Otzen ⁎⁎
⁎
a
Center for Insoluble Protein Structures (inSPIN) and Interdisciplinary Nanoscience Center (iNANO), Department of Molecular Biology and Genetics, Aarhus University,
Gustav Wieds Vej 10C, DK-8000 Aarhus, Denmark
b
Center for Insoluble Protein Structures (inSPIN), Interdisciplinary Nanoscience Center (iNANO), Department of Chemistry, Aarhus University, Langelandsgade 140, DK-8000 Aarhus, Denmark
Interdisciplinary Nanoscience Center (iNANO), Aarhus University, Ny Munkegade 118, DK-8000 Aarhus, Denmark
Department of Biomedicine, Aarhus University, Wilhelm Meyers Allé 4, DK-8000 Aarhus, Denmark
c
d
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 10 August 2011
Received in revised form 14 October 2011
Accepted 24 October 2011
Available online 29 October 2011
The well-ordered cross β-strand structure found in amyloid aggregates is stabilized by many different side
chain interactions, including hydrophobic interactions, electrostatic charge and the intrinsic propensity to
form β-sheet structures. In addition to the side chains, backbone interactions are important because of the
regular hydrogen-bonding pattern. β-Sheet breaking peptide analogs, such as those formed by N-
methylation, interfere with the repetitive hydrogen bonding pattern of peptide strands. Here we test back-
bone contributions to fibril stability using analogs of the 6–10 residue fibril core of human islet amyloid poly-
peptide, a 37 amino acid peptide involved in the pathogenesis of type II diabetes. The Phe–Gly peptide bond
has been replaced by a hydroxyethylene or a ketomethylene group and the nitrogen-atom has been methylated.
In addition, we have prepared peptoids where the side chain is transferred to the nitrogen atom. The backbone
turns out to be extremely sensitive to substitution, since only the minimally perturbed ketomethylene analog
(where only one of the −NH− groups has been replaced by −CH2−) can elongate wildtype fibrils but cannot
fibrillate on its own. The resulting fibrils displayed differences in both secondary structure and overall morphology.
No analog could inhibit the fibrillation of the parent peptide, suggesting an inability to bind to existing fibril sur-
faces. In contrast, side chain mutations that left the backbone intact but increased backbone flexibility or removed
stabilizing side-chain interactions had very small effect on fibrillation kinetics. We conclude that fibrillation is very
sensitive to even small modifications of the peptide backbone.
Keywords:
Amylin
Peptide analog
Amyloid formation
Seeding
Critical aggregation concentration
© 2011 Elsevier B.V. All rights reserved.
1. Introduction
Factors influencing protein fibrillation include protein hydropho-
bicity, electrostatic charge and the propensity to form secondary
Amyloid aggregates mainly composed of fibrillated protein are as-
sociated with a wide variety of diseases such as Alzheimer's and Par-
kinson's disease and type II diabetes [1–3]. The term amyloid is
defined as an in vivo deposited material which displays affinity for
the dye Congo Red, characteristic fibrillar electron microscopy appear-
ance and a specific X-ray diffraction pattern [4]. In addition, a wide
range of unrelated proteins with no involvement in amyloid diseases
is capable of forming fibrillar aggregates with very similar and well or-
dered fibrillar structure [5]. This has inspired the hypothesis that the
ability to form amyloid aggregates is a generic property of proteins.
structures such as α-helix and β-sheet [6,7]. These have successfully
been combined with extrinsic properties such as pH and ionic
strength allowing prediction of the aggregative propensities of different
amino acid sequences [7–9]. Amyloid fibril structures consist of
hydrogen-bonded β-strands forming ladders, which associate laterally
to make contacts between the sidechains forming a steric zipper
[10,11]. Once formed, the fibrillar structures are stabilized by the very
regular hydrogen-bonding patterns along the β-ladders, as can be
seen from the high level of structural persistence in fibrils, correspond-
ing to only one “misplaced” strand per 30,000 strands [12]. For this rea-
son, interference with this backbone pattern could be expected to
severely compromise fibril integrity. Indeed, one strategy to inhibit fibril
growth has been to use β-breaker peptides which bind to the growing
fibril interface but lack the hydrogen bonding partners that allow further
extension and thus act as “terminator peptides” [13–16]. β-Breaking
ability has been achieved by using α-aminobutyric acid which strongly
Abbreviations: GAVL, decapeptide SNNFGAVLSS; GGIL, decapeptide SNNFGGILSS;
NF6, hexapeptide NFGAIL; SN10, decapeptide SNNFGAILSS; ThT, thioflavin T
⁎
Corresponding author. Tel.: +45 89 42 39 32; fax: +45 86 19 61 99.
⁎⁎ Corresponding author. Tel.: +45 89 42 50 46; fax: +45 86 12 31 78.
E-mail addresses: ts@chem.au.dk (T. Skrydstrup), dao@inano.au.dk (D.E. Otzen).
1570-9639/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.bbapap.2011.10.014

M. Andreasen et al. / Biochimica et Biophysica Acta 1824 (2012) 274–285
275
favors α-helical conformations and which sterically prevents β-sheet
contacts, or by using peptides with modified termini. Another approach
has been to modify the peptide backbone with N-methylations or intro-
duction of bulky groups to sterically hinder the formation of hydrogen
bonds [17–23]. The presence of methyl groups on the backbone amide
prevents the formation of hydrogen bonds with the following peptide,
and hence blocks the formation of amyloid fibrils.
We find that even modest changes to the backbone integrity can
have severe effects on the behavior of these peptides with respect
to fibril formation and potential fibril morphology on a local (degree
of ordering, chain packing) and more global scale (twisting, thick-
ness). Only analog Km with the smallest change in the backbone
(−NH− replaced by −CH2−) was able to elongate fibril seeds, and
gave rise to fibrils with different secondary structure and morphology
compared to the original peptide. No other peptide analog could elon-
gate existing fibril seeds. These highlight a central role for the back-
bone in stabilizing the cross-β structure. Moreover, addition of any
analog to wildtype fibrils was unable to prevent or limit growth of
the fibrils.
Here, we report the effect of a single-site backbone modification of
two peptides of length 6 and 10 residues with the sequences NFGAIL
(hIAPP 22–27, denoted NF6 in this work) and SNNFGAILSS (hIAPP
20–29, denoted SN10). These peptides comprise the fibrillating core
of human Islet Amyloid Polypeptide (hIAPP) [24,25]. hIAPP is a 37
amino acid peptide hormone which constitutes the major part of
the pancreatic amyloid found in patients with type II diabetes
[3,26,27]. Fibrils of both the hexa- and decapeptide of the fibrillating
core of hIAPP have been shown to be cytotoxic toward pancreatic
cell lines and this cytotoxicity and the amyloidogenicity can be
blocked by double N-methylated versions of the peptides [18,19,24].
The fibril structure of the decapeptide has been solved using solid-
state NMR, revealing an anti-parallel hetero zipper with an observed
twist along the axis of the fibril [28,29]. As summarized in Fig. 1, we
have modified the FG amide bond of the hexapeptide by the removal
of the nitrogen atom from the backbone amide and introduction of a
hydroxyethylene group (analog He) as well as by preparing an N-
methylated version (analog Nm). Furthermore, we have synthesized
and examined the retropeptoid version of the decapeptide (analog
Rp). Peptoids are oligomers of N-substituted glycines which are iso-
mers of peptides but resistant toward hydrolysis by proteases
[30,31] and have been found to inhibit the formation of fibrils [32].
A ketomethylene isostere of the decapeptide at the same amide
bond has also been prepared and tested (analog Km). As controls,
we investigated two mutants of the decapeptide, namely
SNNFGAVLSS (GAVL peptide) and SNNFGGILSS (GGIL peptide),
which have similar types of structural changes as the chemical ana-
logs but at the level of side chain interactions. Hydrophobic interac-
tions were tested by removal of a methyl group in the I→V
mutation, and the rotational freedom of the peptide backbone was
examined by the A→G mutation in the same position where the
backbone is modified in the peptide analogs.
2. Materials and methods
2.1. Materials
Chemicals were purchased from Sigma-Aldrich, St. Louis, Missouri
and Iris Biotech GmbH, Marktredwitz, Germany. Analogs He and Km
were prepared as described by us [33].
2.2. Peptide synthesis
2.2.1. General method for fully automated microwave-assisted solid
phase peptide synthesis
The peptide synthesis was performed on a CEM Liberty microwave
assisted peptide synthesizer using Rink amide MBHA resin
(0.45 mmol/g) on a 0.1 mmol scale. The resin was swelled for
15 min in 10 mL DMF using the standard protocol of the peptide syn-
thesizer. The coupling cycles used were the standard single coupling
cycles described below:
Fmoc deprotection: 20% piperidine in DMF (7.0 mL) was added,
and microwave heating was employed for 30 s. After draining,
20% piperidine in DMF (7 mL) was added again and the mixture
was heated under microwave radiation for 3 min. Nitrogen gas agi-
tation was used during heating.
Wash: Washings were performed using 7.0 mL of DMF with nitro-
gen gas agitation and were repeated four times.
NFGAIL (NF6)
Asn-Phe-Gly-Ala-Ile-Leu
O
SNNFGAILSS (SN10)
Ala-Ile-Leu-Ser-Ser-NH2
Ala-Ile-Leu-Ser-Ser-NH2
Analogue He
OH
Ser-Asn-Ans-Phe
NH
Ala-Ile-Leu
O
Asn-Phe
O
O
Analogue Km
Ser-Asn-Ans-Phe
Analogue Nm
O
O
Ala-Ile-Leu
Asn-Phe
N
Other unmodified peptides:
CH3
O
SNNFGAVLSS (GAVL peptide)
SNNFGGILSS (GGIL peptide)
Analogue Rp
H C
CH
N
3
3
OH
O
O
CH
N
O
O
O
O
O
O
H
N
3
2
N
N
N
N
NH
HN
N
NH
N
2
H N
O
O
O
3
O
2
CH
OH
CH
OH
3
Fig. 1. Structure of SN10 and peptide analogs used in the present study. Analog He: NF6 hydroxylethylene isostere, Analog Nm: N-methylated NF6, Analog Rp: SN10 retropeptoid,
Analog Km: SN10 ketomethylene isostere, WT: wild-type SN10.

276
M. Andreasen et al. / Biochimica et Biophysica Acta 1824 (2012) 274–285
(0.75 mmol/g). The following protected unmodified and modified
Peptide coupling: Amino acids were added as a 0.20 M solution in
DMF (2.5 mL, 0.5 mmol), HBTU was added as a 0.50 M solution in
DMF (1.0 mL, 0.5 mmol) and DIPEA was added as a 2.00 M solution
in N-methylpyrrolidone (0.5 mL, 1.0 mmol). Microwave heating
was employed for 5 min giving an end temperature of 80 °C. Nitro-
gen gas agitation was used during heating.
natural amino acids were used: Fmoc-Leu-OH, Fmoc-Ile-OH, Fmoc-
Ala-OH, Fmoc-(N-Me)Gly-OH, Fmoc-Phe-OH, and Fmoc-Asn(Trt)-
OH. The cleavage and deprotection of the peptide was performed by
treating the resin with 2 mL of TFA:H2O:TIPS 95:2.5:2.5 (v/v) for a
total of 5 h (2×2.5 h). The crude peptide was dissolved in H2O and
purified by semi-preparative using a linear gradient of 0–60% solvent
B in solvent A. This provided the title compound (31 mg, 48%, 98% pu-
rity). HRMS C31H49N7O8 [M+H+]; calculated: 648.3721, found:
648.3724.
After completing the final Fmoc-deprotection cycle, the peptidyl
resin was dried down by washing with CH2Cl2 (3×4 mL), then with
MeOH (3×4 mL) and finally dried under reduced pressure. Cleavage
and deprotection of the peptide was performed by treating the resin
with 5 mL of TFA:H2O:TIPS 95:2.5:2.5 (v/v) for 2 h at rt. The cleavage
mixture was concentrated to approximately 1 mL under reduced
pressure, followed by precipitation in cold tert-butyl methyl ether.
Repeated centrifugation, decantation, and trituration with tert-butyl
methyl ether (4 times) followed by lyophilization gave the crude
peptide.
2.2.7. Analog Rp (H-NSer-NSer-NLeu-NIle-NAla-Gly-NPhe-NAsn-NAsn-
NSer-NH2)
This peptoid was synthesized by manual solid-phase organic syn-
thesis using a Rink amide MBHA resin (0.2 mmol) [32]. The synthesis
was performed using a combination of the monomer and submono-
mer approach. Residues mimicking serine, asparagine, phenylalanine,
isoleucine, and leucine were performed using the submonomer ap-
proach, while Fmoc-sarcosine-OH was applied for the synthesis of
the alanine residue by the monomer approach. Submonomer method:
The acylations were performed by addition of bromoacetic acid (4 mL,
0.5 M in DMF, 2.0 mmol) and DIC (2.2 mL, 1.0 M in DMF, 2.2 mmol) to
the resin followed by vigorous shaking 20 °C for 30 min, then the reac-
tion mixture was filtered and the reaction was repeated. Nucleophilic
substitution of bromide was performed by reaction of a primary
amine (4 mL, 1 M in DMSO, 4 mmol) for 1 h at 20 °C. The following
amines were used: benzylamine for NPhe, glycinamide hydrochloride
for NAsn, isobutylamine for NLeu, racemic sec-butylamine for NIle and
TBDMS-protected ethanolamine for NSer. N-methyl morpholine
(2.0 mmol) was added in reactions with glycinamide hydrochloride.
Monomer method: Fmoc-sarcosine-OH (187 mg, 0.60 mmol) and
Fmoc-glycine-OH (178 mg, 0.6 mmol) were attached to the growing
peptoid chain using standard solid phase peptide coupling conditions:
HBTU (220 mg, 0.58 mmol), HOBt (81 mg, 0.6 mmol) and DIPEA
(209 μL, 1.2 mmol) in DMF (4 mL) over 1 h [34]. This was followed by
removal of the Fmoc protection groups by repeated treatment of the
resin with 20% piperidine in DMF (5 min×3). After attachment of the
last residue, the resin was dried down by washing with CH2Cl2
(3×4 mL), then with MeOH (3×4 mL) and finally dried under reduced
pressure. Cleavage and deprotection of the peptide was performed by
treating the resin with 5 mL of TFA:H2O:TIPS 95:2.5:2.5 (v/v) for 2 h.
The cleavage mixture was concentrated to approximately 1 mL under
reduced pressure, followed by precipitation in cold tert-butyl methyl
ether. Repeated centrifugation, decantation, and trituration (4 times)
followed by lyophilization gave the crude peptide. The crude mixture
was dissolved in H2O and purified by semi-preparative HPLC using
a linear gradient of 10–60% solvent B in solvent A. This provided
the title compound (46 mg, 22%). HRMS C47H77N13O15 [M+H+];
calculated: 1050.5584, found: 1050.5585.
2.2.2. RP-HPLC
Analytical RP-HPLC was performed using an Agilent 1100 system
with a Phenomenex Kinetex column (C18, 4.6 mm ID×150 mm,
2.6 μm, 100 Å) operated at a flow rate of 1 mL/min at 25 °C. Semi-
preparative RP-HPLC purification was performed on an Agilent 1200
system with a Agilent Zorbax SB column (C18, 9.8 mm ID×250 mm,
5 μm, 300 Å) operated at a flow rate of 5 mL/min at 25 °C. The solvent
system: A=0.1% TFA in H2O, B=0.1% TFA in acetonitrile.
2.2.3. NF6 peptide (H-NFGAIL-OH)
The peptide was synthesized according to the general method ex-
cept that the synthesis was performed at a 0.25 mmol scale using a
pre-loaded Fmoc-Leu-Wang resin (0.75 mmol/g). The Crude peptide
(116 mg) was dissolved in 10% AcOH and purified employing a linear
gradient of 5–60% solvent B in solvent A, yielding the title compound
in a purity of 94.5%. tR =16.73 min (Agilent C18 Stablebond, over
50 min). MALDI-TOF MS: C30H47N7O8 [M+Na+]; calculated 656.34,
found 656.39.
2.2.4. SN10 peptide (H-SNNFGAILSS-NH2)
The peptide was synthesized according to the general method. The
Crude peptide (75 mg) was dissolved in 10% AcOH and purified
employing a linear gradient of 5–40% solvent B in solvent A, yielding
23.3 mg (22%) of the title compound in
tR =9.71 min (Phenomenex Kinetic, over 15 min). MALDI-TOF MS:
43H69N13O15 [M+Na+]; calculated 1030.49, found 1030.50.
a purity of 90.2%.
C
2.2.5. GAVL peptide (H-SNNFGAVLSS-NH2)
The peptide was synthesized according to the general method. The
Crude peptide (94 mg) was dissolved in 10% AcOH and purified employ-
ing a linear gradient of 5–50% solvent B in solvent A, yielding 34.5 mg
(35%) of the peptide in a purity of 91.7%. tR=7.64 min (Phenomenex Ki-
netic, over 15 min). MALDI-TOF MS: C42H67N13O15 [M+K+]; calculated
1032.45, found 1032.55.
2.3. Preparation of samples
All decapeptides (SN10, mutants and analogs) were dissolved in
DMSO to a concentration of 25 mM. Hexapeptides (NF6 and analogs)
were dissolved in DMSO to a concentration of 1000 mM. All peptides
were subsequently diluted 1:100 into 50 mM HEPES buffer pH 7.2
and filtered through a 0.2 μm filter unit. Seeds of preformed fibrils
were produced by centrifugation of fibrils for 30 min at 13,000 rpm
in a tabletop centrifuge. The supernatant was removed and the fibrils
were re-suspended in 50 mM HEPES buffer pH 7.2 to a concentration
of 10 mM for NF6 fibrils and 0.25 mM for SN10 fibrils. Seeds of pre-
formed fibrils were produced by sonication for 10 s at 30% amplitude
with a Bendelin Sonopuls HD 2070 sonication probe (Buch & Holm).
For all seeded fibrillation experiments, 5% preformed fibril seeds
were used to seed the next generation of fibrils.
2.2.5.1. GGIL peptide (H-SNNFGGILSS-NH2). The peptide was synthe-
sized according to the general method. The Crude peptide (90 mg)
was dissolved in 10% AcOH and purified employing a linear gradient
of 5–50% solvent B in solvent A, yielding 33.8 mg (33%) of the title
compound in a purity of 88.1%. tR =8.26 min (Phenomenex Kinetic,
over 15 min). MALDI-TOF MS: C42H67N13O15 [M+Na+]; calculated
1016.48, found 1016.70.
2.2.6. Analog Nm (N-methylated derivative L-Asparaginyl-L-phenylalanyl-
(N-methyl-L-glycyl)-L-alanyl-L-isoleucyl-L-leucinamide)
H-NF(N-Me)GAIL-OH, was synthesized according to the general
method except pre-loaded Fmoc-Leu-Wang resin was employed

M. Andreasen et al. / Biochimica et Biophysica Acta 1824 (2012) 274–285
277
2.4. Fibrillation of peptides and analogs
accessory with a continuous flow of N2 gas. All samples were dried
with N2 gas. 64 interferograms were accumulated at a spectral resolu-
tion of 2 cm−1. Following self deconvolution, the intensity was mod-
eled by Gaussian curve fitting using the OPUS 5.5 software. Peak
positions were assigned where the second order derivative had local
minima.
Thioflavin T (ThT) was added to the protein solution to a final con-
centration of 40 μM and the protein solution was transferred to a 96-
well black Costar polystyrene microtiter plate, sealed to prevent evap-
oration and placed in an Infinite M200 plate reader (Tecan Nordic AB).
When fibril seeds were present, these were added immediately before
the plate was sealed. The plate was incubated at 37 °C and the ThT fluo-
rescence (excitation 450 nm, emission 482 nm) was measured every
5 min with 3 min shaking between each reading.
2.10. X-ray fiber diffraction and analysis
Fiber diffraction specimens were prepared on a stretch frame
using 5 μL fibril suspensions of approximately 2.5 mg/mL. Diffraction
data fibers were collected at room temperature using a fixed wave-
length of 0.8015 Å on a 165 mm MAR-Research CCD detector (Ham-
burg, Germany) on the beamline X13 (EMBL-outstation at DESY,
Hamburg, Germany) using exposure times from 5 to 45 s. The sample
to detector distance was set at 300 mm. The collected diffraction pat-
terns were centered, background corrected by local box averaging
and 60° radial scans of meridional and equatorial reflections were
obtained using Clearer v2.0 for Java v1.5 [35].
2.5. Protein concentration analysis
The concentration of peptides in the supernatant after fibrillation
and centrifugation for 30 min at 13,000 rpm in a tabletop centrifuge
was analyzed using the BCA Protein Assay Reagent kit (Thermo Fisher
Scientific) according to the manufacturers' recommendations. A stan-
dard curve was made from a stock solution of the parent peptides
with known concentration. The concentration of NF6 was also exam-
ined by absorbance at 220 nm on a Nanodrop spectrophotometer
ND1000 (Saveen Werner), also using a standard curve made from a
stock solution with known concentration.
2.11. Solid-state NMR spectroscopy
The samples for solid-state NMR spectroscopy were prepared by
centrifugation of fibril suspensions, removing the supernatant, and
transferring the fibrils to 4 mm (outer diameter) rotors. The samples
contained approximately 4 mg (SN10), 9 mg (GAVL) and 11 mg
(GGIL) of fibrillated peptide. Natural abundance 13C solid state NMR
spectra were acquired using 1H–13C cross polarization in combination
with 12 kHz magic angle spinning (CP/MAS) applying 23,000 scans
per spectrum and repetition delays of 3 s.
2.6. Atomic force microscopy (AFM)
Aliquots of 5 μL fibril solution were deposited onto freshly cleaved
mica surface, air dried for 5 min and finally dried with N2. All images
were captured using a Nanoscope V MultiMode SPM (Veeco Instru-
ments) under ambient conditions. Ultrasharp silicon cantilevers
with a typical resonance frequency of 300 kHz, a spring constant of
26 N/m and
a normal tip radius of 7 nm (triangular, OMCL-
AC160TS-E3, Olympus) were used. AFM imaging was performed in
tapping mode at scan frequency of 1 Hz with minimal loading forces
of 100 pN applied and optimized feedback parameters. The resolution
of all AFM images shown is 512×512 pixels per image. The images
were flattened and analyzed automatically using the Scanning Probe
Image Processor software (SPIPTM, Image Metrology ApS, version
5.1.3).
3. Results
3.1. Fibrillation of IAPP-derived peptides is very sensitive to backbone
modification
We follow the kinetics of fibril formation by monitoring the in-
crease in fluorescence of the fibril-binding dye ThT. Fibrillation of
the parent peptides NFGAIL (NF6) and SNNFGAILSS (SN10) followed
a nucleation dependent pathway. This can be seen in two ways. First-
ly, there is a characteristic lag phase in the accumulation of the ThT
signal (Fig. 2A), which is generally attributed to the build-up of a critical
amount of elongation-competent nuclei [36]. Secondly, the duration of
this lag phase (the lag-time) was reduced by the addition of preformed
fibril seeds. In the absence of fibril seeds, the lag-time of NF6 was found
to range from 6 to 10 h while the lag-time of SN10 was 5.3 0.3 h. The
presence of 5% fibril seeds was found to completely eliminate the lag-
time for both peptides (Fig. 2B). As expected, the presence of preformed
fibril seeds does not change the secondary structure of the fibrils
formed (Fig. 2C).
2.7. Transmission Electron Microscopy (TEM)
Aliquots of 5 μL of fibril solution were mounted on 400-mesh car-
bon coated, glow discharged nickel grids for 30 s. The grids were
washed with one drop of double distilled water and stained with
three drops of 1% phosphotungstic acid pH 7.2. Samples were
inspected in
a JEOL 1010 transmission electron microscope at
60 keV. Images were obtained using an electron sensitive Olympus
KeenView CCD camera.
2.8. Far-UV circular dichroism (CD) spectroscopy
Analog Km is the only decapeptide analog capable of elongating
existing fibril seeds from the parent peptide (Fig. 2B). Over a 70 h pe-
riod, none of the other peptide analogs display an increase in ThT
fluorescence upon incubation with preformed fibril seeds of the parent
peptide. Far-UV CD spectra of these fibrils show very low ellipticity
similar to monomeric peptide, precluding further analysis (Fig. 3A and
B). Although capable of elongating fibril seeds, analog Km is not capable
of forming fibrils in the absence of seeds of the parent peptide or in the
presence of fibril seeds of Km (Fig. 2B). This indicates that analog Km
needs the scaffold of existing well ordered fibrils in order to form the
proper hydrogen bonding pattern needed for fibril formation. Interest-
ingly, the secondary structure of the fibrils formed by analog Km is dif-
ferent from that of fibrils of SN10 as seen from the FTIR spectra (Fig. 3C).
They contain less β-sheet structure (both amyloid and intermolecular)
and also less β-turn structure, while the content of α-helix like struc-
ture is considerably higher (Table 1).
All samples were spun down in
a tabletop centrifuge at
13,000 rpm for 30 min. Samples with visible pellet were resuspended
in 10 mM HEPES buffer pH 7.2. All samples were subjected to sonica-
tion for 2 s at 60% amplitude with a Bendelin probe and inspected for
the presence of visible aggregates in the cuvette prior to analysis. CD
wavelength spectra from 250 nm to 200 nm with a step size 0.2 nm,
bandwidth of 2 nm and scan speed of 50 nm/min were recorded at
25 °C with a J-810 CD-spectrometer (Jasco) using a 1 mm quartz cu-
vette (Hellma). Five spectra were averaged for each sample and the
buffer spectra were subtracted.
2.9. Fourier transform infra-red (FTIR) spectroscopy
FTIR spectroscopy was performed using a Tensor27 FTIR spec-
trometer (Bruker) equipped with Attenuated Total Reflection

278
M. Andreasen et al. / Biochimica et Biophysica Acta 1824 (2012) 274–285
40
30
20
10
0
1600
1400
1200
1000
800
600
400
200
0
6000
5000
4000
3000
2000
1000
0
SN10, 5% SN10 seeds
Rp, 5% SN10 seeds
Km, 5% SN10 seeds
SN10, 5% SN10 seeds, 10% Rp monomer
SN10, 5% SN10 seeds, 10% Km monomer
SN10 monomer
A
A
SN10
NF6 1
NF6 2
NF6 3
-10
-20
0
5
10
15
210
220
230
240
250
Time (h)
Wavelength (nm)
800
600
400
200
B
NF6, 5% NF6 seeds
He, 5% NF6 seeds
Nm, 5% NF6 seeds
NF6, 5% NF6 seeds, 10% He monomer
NF6, 5% NF6 seeds, 10% Nm monomer
NF6 monomer
B
4
2
NF6, 5% NF6 seeds
He, 5% NF6 seeds
Nm, 5% NF6 seeds
SN10, 5% SN10 seeds
Rp, 5% SN10 seeds
Km, 5% SN10 seeds
Km, no seeds
70
60
50
40
30
20
10
0
Km, 5% Km seeds
0
-2
-4
0
5
10
15
200
210
220
230
240
250
Time (h)
Wavelength (nm)
1,2
1
1,2
1
C
C
0,8
0,6
0,4
0,2
0
0,8
0,6
0,4
0,2
0
NF6, no seeds
NF6, 5% NF6 seeds
NF6, 5% NF6 seeds, 10% He monomer
NF6, 5% NF6 seeds 10% Nm monomer
SN10, 5% SN10 seeds
SN10, 5% SN10 seeds, 10% Rp monomer
SN10, 5% SN10 seeds, 10% Km
Km, 5% SN10 seeds
1700
1680
1660
1640
1620
1600
Wavenumber (cm-1)
1700
1680
1660
1640
1620
1600
Wavenumber (cm-1)
Fig. 2. Fibrillation of NF6, SN10 and analogs. A: Non-seeded fibrillation of SN10 and
NF6. The ThT fluorescence of single fibrillations of NF6 is plotted against the left y-axis
while the ThT fluorescence of a triplicate of SN10 is plotted against the right y-axis. The
gain setting for the experiment with SN10 is higher than that of the rest of the experi-
ments reported in the present study giving rise to higher signals. B: Seeded fibrillation
of SN10, NF6 and analogs, and non-seeded fibrillation of analog Km. In the seeded fibrilla-
tion wt seeds were used of the same residue length as the monomer (i.e. SN10 or NF6). C:
Normalized FTIR signal of fibrils of NF6. The only analog capable of elongating fibril seeds
of parent fibrils is analog Km.
Fig. 3. Analysis of the secondary structure of fibrils by far-UV CD and FTIR. A: Far-UV CD
of fibrils of SN10 analogs. B: Far-UV CD of fibrils of NF6 and NF6 analogs. C: Normalized
FTIR signal of fibrils of SN10 and SN10 analogs.
absence of SN10 fibril seeds (Fig. 4A). The fibrillation time course of
the GGIL peptide is essentially identical in the absence and presence
of SN10 fibrils, while there is a clear decrease in lag time for GAVL.
This indicates that GAVL, but not GGIL, elongates the SN10 fibril
seeds. For both peptide mutants, far-UV CD spectra indicate that the
presence of SN10 seeds does not alter the secondary structure of the
ensuing fibrils. However, both FTIR and far-UV CD spectra show that
the secondary structure of fibrils of both mutants are different from
3.2. Side chain mutations have much smaller effects on fibrillation
In contrast to the backbone-modified peptides, both decapeptide
mutants GGIL and GAVL fibrillate in the presence as well as in the

M. Andreasen et al. / Biochimica et Biophysica Acta 1824 (2012) 274–285
279
Table 1
the far-UV CD spectra when GGIL is present during the fibrillation
(Fig. 4C). This indicates that the mutant can leave an imprint on the
SN10 fibrils formed. Taken together, this indicates even the small
changes in hydrophobic interaction and the rotational freedom
around the peptide backbone can have a significant impact on the
structure of the fibrils formed.
Contents of secondary structural elements of fibrils obtained from deconvolution of
FTIR spectra of fibrils according to references [50,51].
Sample
Amyloid
β-sheet
β-Sheet
β-Turn
α-Helix
Random
coil
SN10, 5% SN10 seeds
SN10, 5% SN10 seeds,
10% Rp monomer
SN10, 5% SN10 seeds
10% Km monomer
D, 5% SN10 seeds
SN10, 5% SN10 seeds
GGIL, 5% SN10 seeds
GAVL, 5% SN10 seeds
SN10, 5% SN10 seeds,
10% GGIL monomer
SN10, 5% SN10 seeds,
10% GAVL monomer
NF6
33.6
20.7
31.4
46
30.5
32.9
0.6
0.2
–
–
Peptides lacking hydrogen bonding potential might act as β-
breakers and thus inhibit fibrillation by blocking elongation. However,
this was not observed for any of the hexa- or decapeptide analogs syn-
thesized for this study. The addition of 10% of peptide analogs or 10%
peptide mutant to the parent peptide during fibrillation did not affect
the ThT fluorescence time curve (Fig. 4D). This might be taken to indi-
cate that the parent peptide fibrillates without any interaction with
the peptide analog or peptide mutant. However, as seen for the GGIL
and GAVL peptides above, the structure of the fibrils formed indicates
that in some instances the peptide analogs and the mutants change
the fibril structure. In the case of NF6, the presence of peptide analogs
does not affect the secondary structure as evident from the FTIR data
(Fig. 2C). This can also be seen for SN10 fibrillated in the presence of an-
alog Rp. But in the case of SN10 both with analog Km and mutants, the
secondary structure of the fibrils is affected by the presence of analogs
and mutants, respectively (Figs. 3C and 4B and Table 1). The fibrils of
SN10 made with 10% of analog Km have a secondary structure which
is in between SN10 fibrils and Km fibrils with respect to both the con-
tent of amyloid β-sheet and α-helix. This indicates that the analog is ca-
pable of being incorporated into fibrils of SN10 made from preformed
SN10 fibril seeds.
21.3
33.3
34.2
10.9
–
18.2
34.5
32.4
36.5
32.8
28
25
29.1
26.4
5.4
6.4
36.2
35.9
35.2
6
29.2
2.5
1.8
0.9
29.2
–
–
–
–
6.6
26.4
20.7
34.3
4.3
11.6
18.7
21.2
20
33.8
33.3
32.4
34.3
35.5
34.2
–
–
–
–
–
–
NF6, 5% SN10 seeds
NF6, 5% SN10 seeds,
10% He monomer
NF6 5% SN10 seeds,
10% Nm monomer
18
32.1
32.5
–
–
that of SN10 fibrils (Fig. 4B and C), particularly for GGIL. A small
change in the secondary structure of SN10 fibrils, involving a
~1.5 nm blue shift of the minimum around 222.5 nm, is also seen in
1500
1,2
B
1
A
GGIL, 5% SN10 seeds
GAVL, 5% SN10 seeds
1000
0,8
0,6
0,4
GGIL, no seeds
GAVL, no seeds
500
SN10, 5% SN10 seeds
GGIL, 5% SN10 seeds
GAVL, 5% SN10 seeds
0,2
SN10, 5% SN10 seeds, 10% GGIL monomer
SN10, 5% SN10 seeds, 10% GAVL monomer
0
0
0
0,5
1
1,5
2
4
6
8
10 12 14
1700
1680
1660
1640
1620
1600
Time (h)
Wavenumber (cm-1)
30
20
10
0
1000
SN10, 5% SN10 seeds
GGIL, 5% SN10 seeds
GAVL, 5% SN10 seeds
SN10, 5% SN10 seeds, 10% GGIL monomer
SN10, 5% SN10 seeds, 10% GAVL monomer
GGIL, no seeds
C
D
800
600
400
200
0
GAVL, no seeds
NF6, 5% NF6 seeds, 10% He monomer
NF6, 5% NF6 seeds, 10% Nm monomer
SN10, 5% SN10 seeds, 10% Rp monomer
SN10, 5% SN10 seeds, 10% Km monomer
-10
-20
210 215 220 225 230 235 240 245 250
Wavelength (nm)
0
5
10
15
Time (h)
Fig. 4. Fibrillation of SN10 mutants and inhibition of fibrillation of the parent peptide in the presence of analogs. A: Fibrillation of SN10 mutants with and w/o fibril seeds. The mu-
tants are capable of fibrillation both with and w/o fibril seeds. B: Normalized FTIR signal of fibrils of SN10 mutants. C: Far-UV CD of fibrils of SN10 mutants. D: Fibrillation of parent
peptide in the presence of analogs. No inhibition of the fibrillation is observed.

280
M. Andreasen et al. / Biochimica et Biophysica Acta 1824 (2012) 274–285
3.3. The critical concentration for fibrillation is very sensitive to peptide
length
A
4.61Å
4.78Å
The critical concentration (cc) for fibrillation to occur can be deter-
mined by centrifugation and subsequent measurements of the con-
centration of peptide in the supernatant. If fibrillation is viewed as a
phase transition between soluble monomer and insoluble fibrils,
this concentration is the initial monomer concentration which
needs to be exceeded for fibrillation to occur, and it is the concentra-
tion of monomers in the solution after fibrillation has ended [37]. The
lower the cc, the more easily the peptide is incorporated into the
growing fibrils [38]. The cc of NF6 and SN10 were determined to be
6.8 0.3 and 0.041 0.001 mM, respectively (Table 2). This more
than 100-fold difference in cc explains why much higher concentra-
tions of NF6 are required for fibrillation compared to SN10. For analog
Km seeded with SN10, the cc was found to be 0.091 0.018 mM, indi-
cating that the backbone modification leads to a relatively modest ~2-
fold destabilization of the fibrils, corresponding to a decrease in fibril
stability by ΔΔGfib =−RTln(0.041/0.091)=−0.48 kcal/mol. The cc of
the two SN10 mutants does not change significantly when 5% SN10 fi-
bril seeds are present (Table 2). This suggests that the slight alteration
in secondary structure caused by the SN10 seeds does not affect the
stability of the fibrils.
8.7-8.8Å
B
0,06
0,04
0,035
0,03
0,025
0,02
0,015
0,01
0,005
0
Km
4.78 Å
SN10
GAVL
GGIL
4.61 Å
4.57Å
0,05
0,04
0,03
0,02
0,01
3.4. Structural analysis of the different fibrils highlights different morphologies
Synchrotron X-ray diffraction patterns and the spacing of the re-
flections were obtained from both partially aligned fibril samples
originating from seeded (Fig. 5A and B) or non-seeded fibrils (data
not shown). X-ray diffraction patterns often yield simple patterns
with 4.7–4.8 Å meridional reflections and ~8–10 Å equatorial reflec-
tions. These arise from molecular spacings corresponding to
hydrogen-bonded β-strands arranged perpendicular to the fiber axis
and the β-sheet distance, respectively [39].
In this study, the alignment of fibers of SN10, the GGIL and GAVL
mutant peptides and analog Km was not sufficient to obtain clearly
resolved meridional and equatorial reflections (Fig. 5B). However, ra-
dial scans of the recorded patterns revealed essentially similar reflec-
tions including a 4.78 Å peak corresponding to the β-strand spacing
and a diffuse reflection of weak intensity at ~8.7–8.8 Å characteristic
for the amyloid cross-β structure. The position of these reflections is
in good agreement with previous work by other groups [39,29]. This
suggests that the changes imposed to the polypeptide chain do not
alter the overall amyloid structure.
The radial scans further revealed subtle differences in the
~4.57–4.61 Å region (Fig. 5B). More specifically, fibrils from SN10
and the GGIL mutant, which showed the same ability to elongate
seeds as SN10 (Fig. 4A), gave rise to a reflection at 4.61 Å. This partic-
ular distance was reduced to ~4.57 Å in GAVL which also showed a
reduced Thioflavin T fluorescence in elongation assays compared to
GGIL (Fig. 4A). The ~4.57–4.61 Å reflection was not detected in sam-
ples of the Km analog seeded with wt fibril seeds. This analog did not
fibrillate on its own but required seeding with SN10, suggesting that a
molecular interaction related to this distance may represent a
0,205 0,21 0,215 0,22 0,225 0,23
Reciprocal distance [Å]
8.7-8.8Å
0,1
0,15
0,2
0,25
0,3
Reciprocal distance [Å]
Fig. 5. Analysis of X-ray fiber diffraction pattern of fibers prepared from seeded SN10,
GAVL, GGIL and analog Km samples. (A) Diffraction pattern of seeded SN10. (B) Radial
scans of diffraction patterns from panel A with the reflection at ~4.78 and 8.7–8.8 Å char-
acteristic for amyloid fibrils (highlighted in the zoomed insert) and subtle differences be-
tween samples in the 4.57–4.61 Å region.
structural motif important for promoting fibrillation in the absence
of preformed seeds. It is not clear which molecular packing corre-
spond to the 4.55–4.61 Å reflections. However, their existence and
variability indicate subtle differences in the packing of the fibrils.
Apart from this lacking reflection in fibrils of analog Km, the different
fibrils show highly similar reflections indicative of a common overall
structure, despite the changes imposed on the polypeptide chain.
3.5. ssNMR spectra confirm that different peptides form different structural
classes of fibrils
Table 2
Critical concentration of fibrillation and corresponding ΔGfib values for the peptides
analyzed.
To further explore variations in the fibril structures at the atomic
level, we acquired one-dimensional solid-state NMR spectra for repre-
sentative fibrils. We note here that 13C natural abundance NMR is not
a high-sensitive technique, and the spectra should by no means be com-
pared with spectra of 13C labeled peptides. More detailed investigations
Sample
Cc (mM)
ΔGfib (kJ/mol)
NF6
SN10
6.8 0.3
−3.1
−6.2
−5.7
−6.0
−5.9
−6.2
−6.1
0.041 0.0005
0.091 0.018
0.061 0.014
0.073 0.002
0.044 0.005
0.052 0.009
Analog Km, 5% SN10 seeds
GGIL peptide
GAVL peptide
GGIL, 5% SN10 seeds
GAVL, 5% SN10 seeds
as determination of complete high-resolution 3D structures require 13
C
and 15N labeling of the peptides in combination with multidimensional
spectra. However, the chemical shift patterns (line splittings etc.) in

M. Andreasen et al. / Biochimica et Biophysica Acta 1824 (2012) 274–285
281
the well resolved Ile, Leu and Ala methyl groups, revealing splitting
into doublets and further doublet-doubling in some of the cases, where-
as putative splitting for other resonances such as Cα could not be con-
firmed using 1D 13C NMR spectra due to spectral overlap. For SN10,
the resonances were assigned previously using standard approaches
based on 2D solid-state NMR spectra [28].We observe a doubling of
the resonances, and for AACβ and IACδ a further doubling is observed
(Fig. 6 top panel) consistent with the original assignment of SN10 as
an anti-parallel ladder hetero zipper class [28]. A related pattern is
found for the GAVL mutant (Fig. 6 middle panel), revealing doubling
of LCδ2 and ACβ resonances and a further splitting of AACβ as for the
wild-type. A likely interpretation is that this mutant adopts the same fi-
bril zipper class as for the wild-type SN10 fibrils. Conversely, for GGIL
(Fig. 6 bottom panel) we observe a distinct splitting into two reso-
nances, which are well resolved for the Ile, Leu and Ala methyl groups,
but in this case no further splitting of the doublets could be confirmed.
The lack of further splitting could indicate a different zipper class for this
mutant. This interpretation is supported by the observation of signifi-
cant changes of chemical shifts for the Ile methyl groups (up to
1.8 ppm for IBCδ) which may be due to different packing in the zipper.
This alternative zipper class could be a parallel ladder-heterozipper ar-
rangement where the side chains of a certain residue point into the zip-
per in one ladder, while the side chains of the same residue point out of
the zipper in the opposite ladder. However, the 1D spectra do not
completely exclude further minor splitting of the signals, which
would be compatible with antiparallel ladder fibril symmetries. Both
of these mutants particularly GGIL, give very homogenous fibrils with
well resolved signals, indicative of a high local order.
SN10
GAVL
peptide
GGIL
peptide
Fig. 6. 1D 13C CP/MAS solid-state NMR spectra of SNNFGAILSS wild type (top) and mu-
tants (mutated residue shown in bold). Assignments are provided along with each
spectrum. Magnification of the methyl region (24–8 ppm) of the spectra is shown in
the right panel. In the magnifications assymmetric pairs of the same carbon are indicated
by A/B subscripts. Additional splitting of the A/B form, if present, is highlighted with red
asterisks.
simple 1D 13C spectra acquired using natural abundance fibril samples
can provide some indications on fibril symmetries and local order as
demonstrated previously by Nielsen et al. [28]. From the spectra of the
SN10, GAVL, and GGIL fibrils in Fig. 6, it is seen that they all feature
line splitting, suggesting that fibrils are not totally symmetric. These
splitting patterns can distinguish between different fibril classes but
not specifically assign fibrils to a unique symmetry class without the
use of multidimensional solid state NMR data, as was done for the
SN10 fibril [28].For all samples, these splitting were most apparent for
In order to analyze the structural features of fibrils further, we
turned to TEM and AFM imaging. The morphology of the fibrils of
SN10 monomers formed in the absence and in the presence of fibril
seeds both show long, un-branched, twisted ribbon-like fibrils
(Figs. 7 and 8 and summarized in Table 3). A similar morphology is
A
B
C
SN10
SN10 +
SN10
seeds
SN10
+ SN10 seeds
+ 10% GGIL monomer
SN10
D E
F
GAVL +
SN10
seeds
+ SN10 seeds
+ 10% GAVL monomer
GGIL +
SN10
seeds
Km + SN10
seeds
G
H
I
GGIL
GAVL
Fig. 7. Morphology of fibrils examined by Transmission Electron Microscopy. Upper panel: A: Fibrils of SN10; B: Fibrils of SN10 formed in the presence of 5% fibril seeds; C: Fibrils of
SN10 formed in the presence of 5% fibril seeds and 10% GGIL. Middle panel: D: Fibrils of SN10 formed in the presence of 5% fibril seeds and 10% GAVL; E: Fibrils of GGIL formed in the
presence of 5% SN10 fibril seeds; F: Fibrils of GAVL formed in the presence of 5% SN10 fibril seeds. Bottom panel: G: Fibrils of GGIL; H: Fibrils of GAVL; I: Fibrils of analog Km formed
in the presence of 5% SN10 fibril seeds. The scale bar represents 200 nm. The morphology of the fibrils is greatly dependent on the peptide fibrillating both in the presence and
absence of preformed fibril seeds.

282
M. Andreasen et al. / Biochimica et Biophysica Acta 1824 (2012) 274–285
SN10 +
SN10
SN10
seeds
+ SN10 seeds
+ 10% GGIL monomer
SN10
GGIL +
SN10
seeds
GAVL +
SN10
seeds
SN10 + SN10
seeds + 10%
GAVL monomer
Km +
SN10 seeds
GAVL
GGIL
Fig. 8. Morphology of fibrils examined by atomic force microscopy: Upper panel: S1: Fibrils of SN10; S2: Fibrils of SN10 formed in the presence of 5% fibril seeds; S3: Fibrils of SN10
formed in the presence of 5% fibril seeds and 10% GGIL. Middle panel: S4: Fibrils of SN10 formed in the presence of 5% fibril seeds and 10% GAVL; S5: Fibrils of GGIL formed in the
presence of 5% SN10 fibril seeds; S6: Fibrils of GAVL formed in the presence of 5% SN10 fibril seeds. Bottom panel: S7: Fibrils of GGIL; S8: Fibrils of GAVL; S9: Fibrils of analog Km
formed in the presence of 5% SN10 fibril seeds. The scale bar represents 2 μm.The insets are the line profiles along the dash lines. For S3, the arrow d highlights the special structure,
which is similar to that seen in the TEM image. For S9, the inset is a zoomed in image.
observed for all SN10 fibrils, even those formed in the presence of
peptide analogs and peptide mutants. This indicated that the SN10
peptide dictates the overall morphology of the fibrils, even though
the mutants or the analogs have been incorporated into the fibrils
and give rise to differences in the secondary structure of the fibrils.
AFM images indicate a fibril height of 10–30 nm. This could indicate
that several protofibrils wind around each other to form the final fi-
brils. The structure indicated by the arrow in Fig. 8 subpanel 3 could
also reflect fibril fragments stacking on top of each other. Fibrils of ana-
log Km made from elongation of preformed SN10 fibril seeds have a
morphology very unlike the morphology of the SN10 fibrils, displaying
shorter and thinner fibrils that appear to bundle together or branch. We
conclude that although analog Km is able to elongate fibril seeds and
hence form the hydrogen bonding pattern needed for fibril formation,
this bonding pattern is different from that formed by the parent peptide
during fibrillation, and this in turn leads to different fibril morphologies.
The long, un-branched ribbon-like morphology is also observed
for fibrils formed from the peptide mutants GGIL and GAVL. However,
fibrils formed from both these mutants lack the twisting of the fibrils.
A possible reason for this is that the stacking pattern of fibrils into rib-
bons is disturbed by the removal of the methyl groups in the mutants,
reducing the strength of hydrophobic interactions at that position in
the peptide and therefore altering the structures formed. In the mu-
tant GGIL it could also be due to the introduction of more rotational
freedom in the backbone by the A→G mutation.
Table 3
Summary of the morphology of fibrils as obtained from TEM and AFM images.
Monomer
Wt seeds
Structurea
Periodicity (nm)
Height (nm)
1
2
3
4
5
6
7
8
9
SN10
SN10
9SN10:1GGIL
9SN10: 1GAVL
GGIL
GAVL
GGIL
GAVL
Analog Km
×
√
√
√
√
√
×
×
√
F, uB, TR
F, uB, TR
F, uB, TR
F, uB, TR
F, uB, S
F, uB, S
F, uB, S
F, uB, S
–
~14
~150
NA
14–20
NA
NA
NA
NA
NA
~26/~12
~13/~9
6–10
~25/~17
20–30
16–20
22–30
17–24
NA
4. Discussion
We have investigated the consequences of interfering with a sin-
gle peptide bond in the structure of an amyloidogenic peptide. The
perturbations can be ranked as GGIL≈GAVLbanalog Kmbanalogs
He, Nm and Rp, in terms of their ability to reduce fibrillation in the
modified peptide. This is consistent with the degree of chemical modi-
fication of the peptide backbone, where the smallest changes are intro-
duced in analog Km and the greatest degree of perturbation can be
Notes:
a
F: fibrils; uB: un-branched; TR: Twisted ribbon like; S: Sheet like.

M. Andreasen et al. / Biochimica et Biophysica Acta 1824 (2012) 274–285
283
1.6
found in analog Rp followed by analogs He and Nm. It could be con-
ceived that the exchange of a NH with a methylene group in the back-
bone as for Km would structurally alter the peptide due to the loss of
orbital overlap of the nitrogen lone pair and the π* orbital of the con-
necting carbonyl and hence loss of the double bond character for this
particular linkage. Yet, this NH to CH2 substitution should nevertheless
impart only a minimal sterical increase for this linkage compared to the
other modifications, since the carbon center, although being completely
sp3 hybridized, only possesses one additional hydrogen atom compared
to that of the wild-type peptide. Due to the absence of the double bond
character in Km at one of the peptide linkages, this analog would dis-
play a higher degree of rotational freedom around this particular bond
and hence greater flexibility. This could explain why Km cannot fibril-
late on its own, but nevertheless easily adapts a conformation similar
to that of SN10 in the fibrillation process with seeding of this wild-
type peptide. On the other hand, the peptide analogs He and Nm
show greater structural difference with either lack of the carbonyl
group or increase sterical demand through the introduction of a methyl
group on the nitrogen of the linkage, respectively. Combined with the
fact that these peptide derivatives are shorter than SN10, this rationalizes
their inability to fibrillate even with fibril seeding. Even though none of
the peptides or mutants tested had the ability to inhibit fibrillation of
the parent peptide, the analogs that could elongate seeds of the parent
peptide were able to alter the fibril structure of the parent peptide upon
incorporation of the analog/mutant into the SN10 fibrils formed. A
similar polymorphism of the fibrils can also be seen for the fibrillating
peptide hormone glucagon which has been found to adapt different fi-
bril structures depending on fibrillation conditions [40–43] and muta-
1.5
1.4
1.3
1.2
1.1
1
SN10 peptide
GGIL peptide
GAVL peptide
0.9
0.8
0
2
4
6
8
10
12
Amino acid position
Fig. 9. Prediction of aggregation propensities of SN10, GGIL and GAVL made with the
Zyggregator software [8,46]. The aggregation propensity of GGIL is higher than the
other peptides examined in the present study.
NF6 peptide can be expected to have fewer hydrogen bonds between
the individual peptide stands in the amyloid β-sheet. However the
0.5 kcal/mol difference in fibril stability between SN10 and analog
Km is significantly less than the energy of a hydrogen bond. This is
less than expected, given that one backbone amide group is removed
from analog Km as compared to SN10. It is possible that rearrange-
ments in the structure of the analog Km-containing fibrils (which
are seen both at the level of secondary structure, apparent loss of
the ~4.6 Å reflection in diffraction patterns as well as in terms of ul-
trastructure) compensate for the loss of this hydrogen bond by addi-
tional van der Waals interactions. The asymmetric nature of a fibril
with anti-parallel hydrogen bonding between the β-strands, means
that the same side chain experiences four different environments in
different peptide molecules (Fig. 10 and Ref. [28]), and thus a single
amino acid substitution can have different effects in different envi-
ronments, and also lead to different structural rearrangements in dif-
ferent locations. All this will have unpredictable stability
consequences. However, some conclusions can be made. The differ-
ence in cc between SN10 and the two mutants translates to only
0.2–0.3 kcal/mol, and may be attributed to loss of hydrophobic inter-
actions by the shortening of the Ala and the Ile respectively by a
methyl group in the two mutants. These losses are very modest com-
pared to the effect of methyl group removal from the core of globular
proteins [48], indicating that the fibril structure does not pack these
methyl groups very efficiently, so that few stabilizing interactions
are lost upon their removal. This is confirmed by the structure of
the peptide in the fibrils, cf. Fig. 10, where each side chain is found
both point into the zipper and out toward the solvent, making
fewer interactions. Furthermore, a loose packing of the side chains
around both Ala and Gly is observed in one of the geometries. Alto-
gether these factors indicate that the removal of a hydrophobic inter-
action will not have as severe an effect as that seen in e.g. globular
proteins.
tions [44]. Glucagon is able to fibrillate over
a wide range of
conditions, because the backbone is intrinsically disposed toward fibril-
lation, reducing the fibrillation challenge to one of finding the structure
that fits the current fibrillation conditions. The same effects can be ob-
served for a 20 residue fragment of hIAPP when the amino acid is re-
versed and scrambled. This still leads to formation of amyloid like
fibrils but the structure and properties of the aggregates differ from
those of the original peptide [45]. In the present study all the analogs ex-
cept analog Km lead to the fibrils being too destabilized by the modifica-
tion to allow elongation to occur. This indicates that we have
“overstepped the mark”, so that the analogs are not just unable to elon-
gate existing fibrils, but they cannot even bind strongly enough to the fi-
brils to prevent binding of the non-analog peptides.
In contrast, the two side-chain mutants of SN10 (GGIL and GAVL),
which were designed to mimic the increased backbone flexibility in
the three analogs He, Nm and Rp and probe the effect of removing a
methyl group, were able to form fibrils. Further, the kinetics of fibril-
lation of GGIL was not affected by the presence of fibril seeds of the wt
peptide SN10. This could be ascribed to a higher aggregation propen-
sity of this peptide as compared to the wt peptide. Gratifyingly, the
aggregation propensity prediction software Zyggregator [8,46]
found GGIL to have en aggregation propensity which exceeded the
aggregation propensities of the two other peptides with approxi-
mately 0.2 units for the majority of the amino acid residues of the se-
quence (Fig. 9). Although GAVL has a slightly higher aggregation
propensity than SN10 for all residues but the initial 3 (SNN), these
two peptides are in general very similar with regards to aggregation
propensity. Solid state NMR, FTIR and far-UV CD data also concur
that GAVL (a side chain modification) is closer in structure to SN10
than GGIL (an increase in backbone flexibility as well as side chain
modification). Thus, the two mutants promote the amyloidogenicity
in a ranking order reflecting conformational perturbation of the back-
bone, while chemical backbone modifications either reduce it or
completely eliminate it.
Peptoids have also been suggested as inhibitors of protein–protein
interactions including fibrillation [49]. Furthermore they have the ad-
vantage of being protease resistant [31]. However, the results
obtained for analog Rp, the retropeptoid version of SN10, are in accor-
dance with the previous observation that the SN10 retropeptoid was
not capable of forming amyloid structures on its own [32]. The limited
inhibitory effect of the retropeptoid in a 1:1 mixture with SN10 could
explain why no inhibitory effect is seen for analog Rp in this study
where the ratio of SN10 to retropeptoid is 10:1. Previously developed
β-breaker peptide analogs of SN10 comprise double backbone N-
methylation which blocks the hydrogen-bonding pattern between
The difference in critical aggregation concentration between SN10
and NF6 is more than 100-fold, corresponding to a difference in fibril-
lation energy of 3.0 kcal/mol. This value is in the range of several hy-
drogen bonds [47], in good agreement with the fact that the shorter

284
M. Andreasen et al. / Biochimica et Biophysica Acta 1824 (2012) 274–285
Side view
View perpendicular
to fibril axis
Fig. 10. Space fill model of the structure of the NFGAILS fragment of SN10 fibrils [28] showing the side-chain stacking and interactions between individual strands of the hetero zipper. The
fibril is seen in a side view and a view perpendicular to the fibril axis. In the side view the fibril axis progresses from the top to the bottom. The color coding is C-atoms: green, N-atoms:
blue, O-atoms: red, H-atoms: gray. The Cα of Gly is shown in yellow. A loose packing around Gly is seen in every second strand of the fibril.
[7] F. Chiti, M. Stefani, N. Taddei, G. Ramponi, C.M. Dobson, Rationalization of the ef-
individual peptides by removal of the hydrogen donor site found in
the backbone amid-groups [18,19]. Even though no inhibitory effect
fects of mutations on peptide and protein aggregation rates, Nature 424 (2003)
805–808.
is seen for the N-methylated variant of NF6 on the fibrillation of the par-
ent peptide, this variant is also not capable of elongating fibril seeds.
This indicates that the introduction of a backbone N-methylation affects
the ability of the NF6 peptide to form the hydrogen bonds needed to
form amyloid fibrils. Furthermore, the inhibition of fibrillation of the
parent peptide by addition of N-methylated analog required a 1:1
ratio of analog and parent peptide, which is a ~5 times higher fraction
of the analog than what was used in the present study. The N-
methylated analogs used in earlier studies were not able to inhibit the
fibrillation of full length hIAPP [19]. These results are in good agreement
with the results presented here.
The results presented here indicate that the peptide backbone
structure is very important for the formation of amyloid fibrils. This
can be ascribed to the hydrogen bonding pattern of individual pep-
tides or β-strands in the amyloid fibrils being of utmost importance
for stabilizing the amyloid structure. This is consistent with the high
degree of persistence observed for stacking of individual β-strands
in the fibril structure [12]. We have thus shown that amyloid fibrils
do not form spontaneously when backbone integrity is compromised.
[8] G.G. Tartaglia, A.P. Pawar, S. Campioni, C.M. Dobson, F. Chiti, M. Vendruscolo, Pre-
diction of aggregation-prone regions in structured proteins, J. Mol. Biol. 380
(2008) 425–436.
[9] A.M. Fernandez-Escamilla, F. Rousseau, J. Schymkowitz, L. Serrano, Prediction of
sequence-dependent and mutational effects on the aggregation of peptides and
proteins, Nat. Biotechnol. 22 (2004) 1302–1306.
[10] M.R. Sawaya, S. Sambashivan, R. Nelson, M.I. Ivanova, S.A. Sievers, M.I. Apostol,
M.J. Thompson, M. Balbirnie, J.J.W. Wiltzius, H.T. McFarlane, A.Ø. Madsen, R. Riek,
D. Eisenberg, Atomic structures of amyloid cross-β spines reveal varied steric zip-
pers, Nature 447 (2007) 453–457.
[11] R. Nelson, M.R. Sawaya, M. Balbirnie, A.Ø. Madsen, C. Riekel, R. Grothe, D. Eisenberg,
Structure of the cross-beta spine of amyloid-like fibrils, Nature 435 (2005) 773–778.
[12] T.P.J. Knowles, J.F. Smith, A. Craig, C.M. Dobson, M.E. Welland, Spatial persistence
of angular correlations in amyloid fibrils, Phys. Rev. Lett. 96 (2006) 238301.
[13] C. Soto, E.M. Sigurdsson, L. Morelli, R.A. Kumar, E.M. Castano, B. Frangione, Beta-
sheet breaker peptides inhibit fibrillogenesis in a rat brain model of amyloidosis:
implications for Alzheimer's therapy, Nat. Med. 4 (1998) 822–826.
[14] G.M. Musso, C.C. Arico-Muendel, H.W. Benjamin, A.M. Hundal, J.-J. Lee, J. Chin, M.
Kelley, J. Wakefield, N.J. Hayward, S.M. Molineau, Modified-peptide inhibitors of
amyloid β-peptide polymerization, Biochemistry 38 (1999) 6791–6800.
[15] S. Gilead, E. Gazit, Inhibition of amyloid fibril formation by peptide analogs mod-
ified with α-aminoisobutyric acid, Angew. Chem. Int. Ed. 116 (2004) 4133–4136.
[16] L.O. Tjernberg, J. Naslund, F. Lindqvist, J. Johansson, A.R. Karlstrom, J. Thyberg, L.
Terenius, C. Nordstedt, Arrest of beta-amyloid fibril formation by a pentapeptide
ligand, J. Biol. Chem. 271 (1996) 8545–8548.
[17] L.M. Yan, M. Tatarek-Nossol, A. Velkova, A. Kazantzis, A. Kapurniotu, Design of a
mimic of nonamyloidogenic and bioactive human islet amyloid polypeptide
(IAPP) as nanomolar affinity inhibitor of IAPP cytotoxic fibrillogenesis, Proc.
Natl. Acad. Sci. U. S. A. 103 (2006) 2046–2051.
Acknowledgements
[18] A. Kapurniotu, A. Schmauder, K. Tenidis, Structure-based design and study of
non-amyloidogenic, double N-methylated IAPP amyloid core sequences as inhibi-
tors of IAPP amyloid formation and cytotoxicity, J. Mol. Biol. 315 (2002) 339–350.
[19] M. Tatarek-Nossol, L.M. Yan, A. Schmauder, K. Tenidis, G. Westermark, A. Kapur-
niotu, Inhibition of hIAPP amyloid-fibril formation and apoptotic cell death by a
designed hIAPP amyloid- core-containing hexapeptide, Chem. Biol. 12 (2005)
797–809.
[20] N. Kokkoni, K. Stott, H. Amijee, J.M. Mason, A.J. Doig, N-Methylated peptide inhibitors
of beta-amyloid aggregation and toxicity. Optimization of the inhibitor structure,
Biochemistry 45 (2006) 9906–9918.
This work is supported by the Danish National Research Founda-
tion (inSPIN). The authors would like to thank Dr Louise Serpell for
helpful discussions and Dr. Michele Vendruscolo for use of the Zyg-
gregator programme. We acknowledge EMBL-Hamburg for its provi-
sion of synchrotron facilities for X-ray diffraction studies and we
thank Dr. Victor Lamzin and Johanna Kallio for their assistance in
using the X13 beamline.
[21] D.T. Rijkers, J.W. Hoppener, G. Posthuma, C.J. Lips, R.M. Liskamp, Inhibition of amy-
loid fibril formation of human amylin by N-alkylated amino acid and alpha-
hydroxy acid residue containing peptides, Chemistry 8 (2002) 4285–4291.
[22] R.C. Elgersma, T. Meijneke, G. Posthuma, D.T. Rijkers, R.M. Liskamp, Self-assembly
of amylin(20–29) amide-bond derivatives into helical ribbons and peptide nano-
tubes rather than fibrils, Chemistry 12 (2006) 3714–3725.
[23] D.J. Gordon, K.L. Sciarretta, S.C. Meredith, Inhibition of beta-amyloid(40) fibrillo-
genesis and disassembly of beta-amyloid(40) fibrils by short beta-amyloid con-
geners containing N-methyl amino acids at alternate residues, Biochemistry 40
(2001) 8237–8245.
References
[1] J.W. Kelly, The alternative conformations of amyloidogenic proteins and their
multi-step assembly pathways, Curr. Opin. Struct. Biol. 8 (1998) 101–106.
[2] J.C. Rochet, P.T. Lansbury Jr., Amyloid fibrillogenesis: themes and variations, Curr.
Opin. Struct. Biol. 10 (2000) 60–68.
[3] A. Clark, G.J. Cooper, C.E. Lewis, J.F. Morris, A.C. Willis, K.B. Reid, R.C. Turner, Islet
amyloid formed from diabetes-associated peptide may be pathogenic in type-2
diabetes, Lancet 2 (1987) 231–234.
[4] P. Westermark, M.D. Benson, J.N. Buxbaum, A.S. Cohen, B. Frangione, S. Ikeda, C.L.
Masters, G. Merlini, M.J. Saraiva, J.D. Sipe, A primer of amyloid nomenclature, Amyloid
14 (2007) 179–183.
[24] K. Tenidis, M. Waldner, J. Bernhagen, W. Fischle, M. Bergmann, M. Weber, M.L.
Merkle, W. Voelter, H. Brunner, A. Kapurniotu, Identification of a penta- and hex-
apeptide of islet amyloid polypeptide (IAPP) with amyloidogenic and cytotoxic
properties, J. Mol. Biol. 295 (2000) 1055–1071.
[25] P. Westermark, U. Engstrom, K.H. Johnson, G.T. Westermark, C. Betsholtz, Islet
amyloid polypeptide: pinpointing amino acid residues linked to amyloid fibril
formation, Proc. Natl. Acad. Sci. U. S. A. 87 (1990) 5036–5040.
[26] E.L. Opie, The relation of diabetes mellitus to lesions of the pancreas. Hyaline de-
generation of the islands of Langerhans, J. Exp. Med. 5 (1901) 527–540.
[5] C.M. Dobson, Protein misfolding, evolution and disease, Trends Biochem. Sci. 24
(1999) 329–332.
[6] K.F. DuBay, A.P. Pawar, F. Chiti, J. Zurdo, C.M. Dobson, M. Vendruscolo, Prediction
of the absolute aggregation rates of amyloidogenic polypeptide chains, J. Mol.
Biol. 341 (2004) 1317–1326.

M. Andreasen et al. / Biochimica et Biophysica Acta 1824 (2012) 274–285
285
[27] G.J. Cooper, A.C. Willis, A. Clark, R.C. Turner, R.B. Sim, K.B. Reid, Purification and
characterization of a peptide from amyloid-rich pancreases of type 2 diabetic pa-
tients, Proc. Natl. Acad. Sci. U. S. A. 84 (1987) 8628–8632.
[28] J.T. Nielsen, M. Bjerring, M. Jeppesen, R.O. Pedersen, J.M. Pedersen, K.L. Hein, T.
Vosegaard, T.S. Skrydstrup, D.E. Otzen, N.C. Nielsen, Unique identification of su-
pramolecular structures in amyloid fibrils by solid-state NMR, Angew. Chem.
Int. Ed. 48 (2009) 2118–2121.
[38] A. Peim, P. Hortschansky, T. Christopeit, V. Schroeckh, W. Richter, M. Fandrich,
Mutagenic exploration of the cross-seeding and fibrillation propensity of Alzhei-
mer's beta-amyloid peptide variants, Protein Sci. 15 (2006) 1801–1805.
[39] M. Sunde, L. Serpell, M. Bartlam, P. Fraser, M. Pepys, C. Blake, Common core structure
of amyloid fibrils by synchrotron X-ray diffraction, J. Mol. Biol. 273 (1997) 729–739.
[40] J.S. Pedersen, C.B. Andersen, D.E. Otzen, Institutional nonconformism: the many
levels of glucagon polymorphism, FEBS J. 277 (2010) 4591–4601.
[29] J. Madine, E. Jack, P.G. Stockley, S.E. Radford, J.C. Serpell, D.A. Middleton, Structural
insights into the polymorphism of amyloid-like fibrils formed by region 20–29 of
amylin revealed by solid-state NMR and X-ray fiber diffraction, J. Am. Chem. Soc.
130 (2008) 14990–15001.
[41] C.B. Andersen, M.R. Hicks, B. Vandahl, H. Rahbek-Nielsen, H. Thøgersen, I.B.
Thøgersen, J.J. Enghild, L.C. Serpell, C. Rischel, D.E. Otzen, Glucagon fibril polymor-
phism reflects differences in protofilament backbone structure, J. Mol. Biol. 397
(2010) 932–946.
[30] R.J. Simon, R.S. Kania, R.N. Zuckermann, V.D. Huebner, D.A. Jewell, S. Banville, S. Ng, L.
Wang, S. Rosenberg, C.K. Marlowe, D.C. Spellmeyer, R. Tans, A.D. Frankel, D.V. Santi,
F.E. Cohen, P.A. Bartlett, Peptoids: a modular approach to drug discovery, Proc. Natl.
Acad. Sci. U. S. A. 89 (1992) 9367–9371.
[31] S.M. Miller, R.J. Simon, S. Ng, R.N. Zuckermann, J.M. Kerr, W.H. Moos, Comparison
of the proteolytic susceptibilities of homologous L-amino acid, D-amino acid, and
N-substituted glycine peptide and peptoid oligomers, Drug Dev. Res. 35 (1995)
20–32.
[32] R.C. Elgersma, G.E. Mulder, J.A. Kruijtzer, G. Posthuma, D.T. Rijkers, R.M. Liskamp,
Transformation of the amyloidogenic peptide amylin(20–29) into its corresponding
peptoid and retropeptoid: access to both an amyloid inhibitor and template for self-
assembled supramolecular tapes, Bioorg. Med. Chem. Lett. 17 (2007) 1837–1842.
[33] D. Mittag, D.E. Otzen, N.C. Nielsen, T.S. Skrydstrup, Synthesis of a ketomethylene
isostere of the fibrillating peptide SNNFGAILSS, J. Org. Chem. 74 (2009)
7955–7957.
[34] W.C. Chen, P.D. White, Fmoc Solid Phase Peptide Synthesis, Oxford University
Press, Oxford, 2000.
[35] S. Makin, P. Sikorski, L. Serpell, CLEARER: a new tool for the analysis of X-ray fibre
diffraction patterns and diffraction simulation from atomic structural models, J.
Appl. Crystallogr. 40 (2007) 966–972.
[36] A.M. Morris, M.A. Watzky, R.G. Finke, Protein aggregation kinetics, mechanism,
and curve-fitting: a review of the literature, Biochim. Biophys. Acta 1794 (2009)
375–397.
[37] J.T. Jarrett, P.T. Lansbury Jr., Seeding “one-dimensional crystallization” of amyloid:
a pathogenic mechanism in Alzheimer's disease and scrapie? Cell 73 (1993)
1055–1058.
[42] J.S. Pedersen, D. Dikov, J.L. Flink, H.A. Hjuler, G. Christiansen, D.E. Otzen, The
changing face of glucagon fibrillation: structural polymorphism and conforma-
tional imprinting, J. Mol. Biol. 355 (2006) 501–523.
[43] J.S. Pedersen, D.E. Otzen, Amyloid — a state in many guises: survival of the fittest
fibril fold, Protein Sci. 17 (2008) 1–9.
[44] J.S. Pedersen, D. Dikov, D.E. Otzen, N- and C-terminal hydrophobic patches are in-
volved in fibrillation of glucagon, Biochemistry 45 (2006) 14503–14512.
[45] R. Sabate, A. Espargaro, N.S. de Groot, J.J. Valle-Delgado, X. Fernandez-Busquets, S.
Ventura, The role of protein sequence and amino acid composition in amyloid
formation: scrambling and backward reading of IAPP amyloid fibrils, J. Mol.
Biol. 404 (2010) 337–352.
[46] G.G. Tartaglia, M. Vendruscolo, The Zyggregator method for predicting protein ag-
gregation propensities, Chem. Soc. Rev. 37 (2008) 1395–1401.
[47] A.R. Fersht, J.P. Shi, J. Knill-Jones, D.M. Lowe, A.J. Wilkinson, D.M. Blow, P. Brick, P.
Carter, M.M.Y. Waye, G. Winter, Hydrogen bonding and biological specificity ana-
lysed by protein engineering, Nature 314 (1985) 235–238.
[48] A.R. Fersht, S.E. Jackson, L. Serrano, Protein stability: experimental data from protein
engineering, Phil. Trans. R. Soc. Lond. 345 (1993) 141–151.
[49] R. Ruijtenbeek, J.A. Kruijtzer, W. van de Wiel, M.J. Fischer, M. Fluck, F.A. Redegeld,
R.M. Liskamp, F.P. Nijkamp, Peptoid — peptide hybrids that bind Syk SH2 domains
involved in signal transduction, ChemBioChem 2 (2001) 171–179.
[50] J. Kong, S. Yu, Fourier transform infrared spectroscopic analysis of protein secondary
structures, Acta Biochim. Biophys. Sin. (Shanghai) 39 (2007) 549–559.
[51] G. Zandomeneghi, M.R. Krebs, M.G. McCammon, M. Fandrich, FTIR reveals struc-
tural differences between native beta-sheet proteins and amyloid fibrils, Protein
Sci. 13 (2004) 3314–3321.