Influence of sequential modifications and carbohydrate variations in synthetic AFGP analogues on conformation and antifreeze activity

Article abstract of DOI:10.1002/chem.201202119

Certain Arctic and Antarctic ectotherm species have developed strategies for survival under low temperature conditions that, among others, consist of antifreeze glycopeptides (AFGP). AFGP form a class of biological antifreeze agents that exhibit the ability to inhibit ice growth in vitro and in vivo and, hence, enable life at temperatures below the freezing point. AFGP usually consist of a varying number of (Ala-Ala-Thr)_n units (n=4-55) with the disaccharide β-D-galactosyl-(1→3)-α-N-acetyl-D-galactosamine glycosidically attached to every threonine side chain hydroxyl group. AFGP have been shown to adopt polyproline II helical conformation. Although this pattern is highly conserved among different species, microheterogeneity concerning the amino acid composition usually occurs; for example, alanine is occasionally replaced by proline in smaller AFGP. The influence of minor and major sequence mutations on conformation and antifreeze activity of AFGP analogues was investigated by replacement of alanine by proline and glycosylated threonine by glycosylated hydroxyproline. The target compounds were prepared by using microwave-enhanced solid phase peptide synthesis. Furthermore, artificial analogues were obtained by copper-catalyzed azide-alkyne cycloaddition (CuAAC): propargyl glycosides were treated with polyproline helix II-forming peptides comprising (Pro-Azp-Pro)_n units (n=2-4) that contained 4-azidoproline (Azp). The conformations of all analogues were examined by circular dichroism (CD). In addition, microphysical analysis was performed to provide information on their inhibitory effect on ice recrystallization. Out of the cold: The antifreeze activity of AFGP (antifreeze glycoprotein, see figure) analogues strongly depends on the amino acid composition. Synthetic, near-native compounds (e.g., with proline replacing alanine) retain the effect, whereas others exhibit reduced or abolished influence on ice recrystallization, despite having high propensity to adopt polyproline II helical conformation, as proposed for the native AFGP. Copyright

Full text of DOI:10.1002/chem.201202119

FULL PAPER
DOI: 10.1002/chem.201202119
Influence of Sequential Modifications and Carbohydrate Variations in
Synthetic AFGP Analogues on Conformation and Antifreeze Activity
Lilly Nagel,[a] Carsten Budke,[b] Roman S. Erdmann,[c] Axel Dreyer,[b]
Helma Wennemers,[c] Thomas Koop,[b] and Norbert Sewald*[a]
Abstract: Certain Arctic and Antarctic
ectotherm species have developed
strategies for survival under low tem-
perature conditions that, among others,
consist of antifreeze glycopeptides
(AFGP). AFGP form a class of biolog-
ical antifreeze agents that exhibit the
ability to inhibit ice growth in vitro and
in vivo and, hence, enable life at tem-
peratures below the freezing point.
AFGP usually consist of a varying
number of (Ala-Ala-Thr)n units (n=4–
55) with the disaccharide b-d-galacto-
syl-(1!3)-a-N-acetyl-d-galactosamine
glycosidically attached to every threo-
nine side chain hydroxyl group. AFGP
have been shown to adopt polyproli-
ne II helical conformation. Although
this pattern is highly conserved among
different species, microheterogeneity
concerning the amino acid composition
usually occurs; for example, alanine is
occasionally replaced by proline in
smaller AFGP. The influence of minor
and major sequence mutations on con-
formation and antifreeze activity of
AFGP analogues was investigated by
replacement of alanine by proline and
glycosylated threonine by glycosylated
hydroxyproline. The target compounds
were prepared by using microwave-en-
hanced solid phase peptide synthesis.
Furthermore, artificial analogues were
obtained by copper-catalyzed azide–
alkyne cycloaddition (CuAAC): prop-
argyl glycosides were treated with poly-
proline helix II-forming peptides com-
prising (Pro-Azp-Pro)n units (n=2–4)
that contained 4-azidoproline (Azp).
The conformations of all analogues
were examined by circular dichroism
(CD). In addition, microphysical analy-
sis was performed to provide informa-
tion on their inhibitory effect on ice re-
crystallization.
Keywords: bioorganic chemistry ·
circular dichroism · glycopeptides ·
ice recrystallization
chemistry
· microwave
Introduction
differ in the range 5–35 mgmLꢀ1.[1,2] These glycoproteins are
supposed to directly interact with the ice surface with impli-
cations on ice crystal growth. Besides the inhibition of ice
recrystallization, heterogeneous nucleation is suppressed
and a change in crystal habitus occurs.[3] The non-equilibri-
um hysteresis freezing point (HFP) is depressed, whereas
the equilibrium freezing point (EFP) is changed only mar-
ginally. The difference between the HFP and EFP is refer-
red to as the thermal hysteresis (TH).[1b,4]
In contrast to AFP that are accessible by recombinant
techniques and, hence, have been investigated thoroughly,[5]
AFGP can be obtained only in small amounts as pure sam-
ples from natural sources. Chemical and biochemical ap-
proaches for obtaining homogeneous AFGP are much more
demanding. Consequently, the molecular mechanism of
action and nature of the interface between the proteins and
the ice surface has not yet been fully elucidated.[1b] Recently
AFGP were implicated to disturb long-range hydration dy-
namics of the ice surface.[6]
A variety of ectotherm marine species in polar and subpolar
regions have developed strategies to survive at temperatures
below the colligative freezing point of their body fluids. This
phenomenon relies on regulation of ice nucleation and ice
crystal growth. The sera of such fish contain a high concen-
tration of salts, carbohydrates and amino acids. However, an
additional protective effect leading to a higher freezing
point depression is accomplished by macromolecules such as
antifreeze proteins (AFP) and antifreeze glycoproteins
(AFGP).[1] The latter have been found in blood, body tissues
and the intestines of several Arctic and Antarctic fish at
concentrations that largely vary with the season and can
[a] Dr. L. Nagel, Prof. Dr. N. Sewald
Organic and Bioorganic Chemistry, Bielefeld University
Universitꢀtsstrasse 25, 33615 Bielefeld (Germany)
Fax : (+49)521-106-8094
E-mail: norbert.sewald@uni-bielefeld.de
AFGP are iterative peptides and proteins comprising of a
tripeptide unit (Ala-Ala-Thr)n (n=4–55) in a molecular
mass range between 2.6 and 33.7 kDa. They are glycosylated
with b-d-galactosyl-(1!3)-a-N-acetyl-d-galactosamine at
each threonine hydroxyl group.[1b,4c,7] Interestingly, the same
glycosyl residue is also found in mammals where it is known
as a tumor-associated carbohydrate antigen (T-antigen). Al-
though the primary structure is largely conserved in AFGP,
[b] Dr. C. Budke, Dr. A. Dreyer, Prof. Dr. T. Koop
Physical Chemistry, Bielefeld University
Universitꢀtsstrasse 25, 33615 Bielefeld (Germany)
[c] R. S. Erdmann, Prof. Dr. H. Wennemers
Laboratory for Organic Chemistry, ETH Zꢁrich
Wolfgang-Pauli-Strasse 10, 8093 Zꢁrich (Switzerland)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201202119.
Chem. Eur. J. 2012, 00, 0 – 0
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

some isoforms have been characterized. For instance, ala-
nine occasionally is substituted by proline, particularly in
small AFGP.[8] Threonine sometimes is replaced by arginine
in the Northern notothenoides.[9] Other modifications were
found to be tolerated, too; for example, N-acetylglucos-
amine was discovered as the exclusive carbohydrate compo-
nent in Pleagranna antarcticum.[10] There is evidence that
the antifreeze activity depends on several essential features,
such as free hydroxyl groups and the galacto configuration
of the carbohydrate moieties, the acetamido group in C2 po-
sition, the a-glycosidic bond between the d-galactosamine
and the threonine residue, and the g-methyl group in the
side chain of every threonine residue.[11] The presence of the
second galactosyl residue is not crucial for freezing point de-
pression, but significantly improves the antifreeze activity.[11]
The first synthetic approach to AFGP analogues was ac-
complished by Tsuda and Nishimura who used Ala-Ala-Thr
tripeptide fragments, which were glycosylated with a benzy-
lidene and acetate protected disaccharide fluoride. Subse-
quently, the fully deprotected tripeptide precursors were
polymerized with diphenylphosphoryl azide (DPPA) to give
peptides with molecular masses in the range of 6.0–7.3 kDa
(n=10–12).[12] The same group further elaborated this syn-
thetic strategy and slightly modified the protecting group
approach. Here the benzyl protected disaccharide halide
was coupled to Ala-Thr-Ala tripeptide units, which were
polymerized providing glycopeptides with a molecular mass
of about 6.6 kDa.[13] However, this fragment oligomerization
approach does not allow for variations in sequence and pep-
tide length and homogenous product is obtained only after
tedious separation.
An alternative synthetic route relies on the preparation of
a glycosylated amino acid building block, that is incorporat-
ed by solid phase peptide synthesis.[14] The T-antigen precur-
sor (Gal-GalNAc) could either be generated by coupling of
a carbohydrate to TN-antigen (GalNAc) or the amino acid is
glycosylated with a disaccharide donor. The latter approach
was communicated by us recently.[15] A tert-butyldiphenylsil-
yl (TBDPS) protected galactal was converted into a glycosyl
donor by azidochlorination, and coupled subsequently to or-
thogonally protected threonine forming the desired a-
anomeric product. After conversion of the azide moiety in
C2 position into the acetamido group and changes in the
protecting group scheme the Fmoc-protected and disacchar-
ide-substituted threonine building block was introduced by
microwave-enhanced SPPS.[15] Several further approaches
regarding mutations in primary structure, usage of different
carbohydrate moieties and bond motifs apart from O-glyco-
sides were evaluated to explore the molecular mechanism of
action between AFGP and the ice surface. Recently, we
published the synthesis of peptoids with triazol-linked car-
bohydrate moieties, which showed no significant antifreeze
activity.[16] Peltier et al. reported on the analysis of monosac-
charide-substituted AFGP analogues with alanine residues
being replaced by proline. Both the lack of the second galac-
tosyl component and the sequential variations resulted in
decreased antifreeze activity.[17] We published the synthesis
of AFGP analogues containing three to four (Ala-Ala-Thr)
repeating units substituted by the monosaccharide TN-anti-
gen. Besides naturally occurring AFGP, analogues with
amino acid replacements with glycine, proline and serine
residues instead of alanine were synthesized. The glycopep-
tides with mutated sequence showed less activity than the
(Ala-Ala-Thr)-glycopeptides due to aperiodic turn forma-
tions caused by Pro, Gly and Ser.[18] Miller et al. reported on
the synthesis of azido-functionalized amino acids or pepti-
des, which were glycoconjugated by CuAAC under micro-
wave conditions with a propargylated N-acetylgalactosamine
giving triazole-linked AFGP analogues.[19] However, no re-
crystallization assays were conducted to gain the activity in
ice growth inhibition.
Conformational analysis by using NMR spectroscopy, CD
experiments, Raman and IR spectroscopy suggest that low-
molecular AFGP adopt highly flexible left-handed helices
similar to a polyproline II (PPII) helix.[20] Bush and Feeney
stated that the carbohydrate residues are aligned at the
same side forming the hydrophilic face and the methyl
groups of the alanine and threonine residues represent the
hydrophobic face. Such peptides represent amphipathic mol-
ecules enabling binding to the ice surface with the hydro-
philic side.[11,21] The larger glycopeptides AFGP1–4 are sup-
posed to behave like flexible rods showing a high segmental
mobility.[22]
Results and Discussion
We recently communicated a novel synthetic strategy for
preparation of suitably glycosylated threonine building
blocks either the T- or TN-antigen. It can be incorporated
into peptides by microwave-enhanced SPPS providing
AFGP analogues predefined in length and sequence. We
adapted this approach to different amino acids, such as hy-
droxyproline (Hyp), and other compounds, like propargyl
alcohol. Glycopeptides and glycopeptide analogues contain-
ing these various glycosylated derivatives were prepared.
Alanine residues were replaced by different amino acids,
such as proline. The main target is to associate the modified
structure of these glycopeptides with a change in conforma-
tion and antifreeze activity. Conformational analysis of
these peptides was done by using CD spectroscopy. The an-
tifreeze activity of the synthetic analogues was tested in re-
crystallization assays.
Monosaccharide-substituted AFGP analogues with partial
replacement of alanine by proline as observed in peptides
from natural sources were synthesized. As AFGP are sup-
posed to adopt a polyproline II helix, a series of chimeric
peptides displaying features of PPII helix-forming collagen
(Gly-Pro-Hyp tripeptide repeating unit) was synthesized. A
monosaccharide-substituted hydroxyproline residue was in-
corporated as a precursor into a sequence composed of Gly-
Pro-Hyp
motif.
ACHTUNGTREN(NGNU GalNAc) tripeptide units, mimicking the collagen
&
2
&
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Modified AFGP Analogues
FULL PAPER
As an additional modification, propargyl glycosides were
conjugated by copper catalyzed azide–alkyne cycloaddition
(CuAAC) to oligoprolines containing 4-azidoproline (Azp)
in every third position [(Pro-Azp-Pro)n]. Such azidoproline
containing oligoprolines are known to form PPII helix and
are easily functionalizable.[23]
The synthesis of glycosylated threonine was accomplished
as published previously by us.[15,18,24] This protocol was
adapted for glycosylation of hydroxyproline (Scheme 1).
Scheme 1. Synthesis of the monosaccharide-substituted Hyp building
block for the synthesis of AFGP analogues. a) Fmoc-l-trans-Hyp-OtBu,
Ag2CO3, AgClO4, CH2Cl2/toluene (1:1), room temperature, 40 h, 56%;
b) AcSH/pyridine (2:1), toluene, 508C, 30 min, 54%; c) TFA/TIS/H2O
(95:2.5:2.5), 30 min, 69%.
Scheme 2. Synthesis of the glycosylated propargyl alcohol building block
for the synthesis of triazole-linked AFGP analogues. d) Ag2CO3,
AgClO4, CH2Cl2/toluene (1:1), room temperature, 40 h, 49% (5), 60%
(8); e) AcSH/pyridine (2:1), toluene, 508C, 30 min, 43% (6), 51% (9);
f) cat. NaOMe, MeOH, room temperature, 20 h, 95% (7); g) 1. 1m
TBAF in THF, room temperature, 1 h, 2. cat. NaOMe, MeOH, room tem-
perature, 20 h, 49% (10).
The azidochloride[24] served as Koenigs–Knorr donor and
was coupled to the orthogonally protected amino acids
Fmoc-l-Thr-OtBu and Fmoc-l-trans-Hyp-OtBu. The glyco-
sylated Thr building block was obtained in an anomeric
ratio a/b of 3:1, whereas glycosylated Hyp had an anomeric
ratio a/b of 9:1. Simultaneous azide reduction and N-acety-
lation succeeded smoothly for both amino acid derivatives
and the anomers could be separated on column chromatog-
raphy yielding exclusively a-configured product. After
cleavage of the tBu ester the monosaccharide-substituted
Thr building block was used without further purification in
solid phase peptide synthesis (SPPS). In the case of the
monosaccharide-substituted Fmoc-Hyp-OtBu ester cleavage
with TFA and water both in the absence and presence of
TIS resulted in anomerization to give an a/b ratio of about
3:1. The separation of the diastereomers was achieved by
column chromatography.
The glycosylation step and the subsequent introduction of
the N-acetyl group in C2 position of the propargyl glyco-
sides comprising the mono- and disaccharide succeeded as
mentioned for the Thr and Hyp building blocks
(Scheme 2).[15,24] The anomeric ratio a/b was determined to
be 5:1 for the propargyl monosaccharide derivative and 3:1
for the propargyl disaccharide, whereas the glycosylation of
threonine with a TBDPS protected disaccharide donor
solely gave a-anomeric product due to the steric demand of
the TBDPS protecting group in C6 position of the disacchar-
ide moiety and of Fmoc-l-Thr-OtBu as the glycosyl accept-
or.[15] In the case of the smaller propargyl alcohol the bulky
TBDPS group did not have the ability to direct the glycosy-
lation exclusively to a-configured product. After conversion
of the azide in C2 position into the acetamide the anomers
could be successfully separated by column chromatography.
Cleavage of the acetate groups in the propargyl monosac-
charide was carried out by using catalytic amounts of
sodium methanolate in methanol. In the case of the prop-
argyl disaccharide the TBDPS group was cleaved with
TBAF in THF prior to cleavage of the acetate groups.[25]
The syntheses of the AFGP analogues containing Pro res-
idues or Gly-Pro-Hyp tripeptide units (Scheme 3 and
Scheme 4) were conducted by stepwise microwave-assisted
SPPS.[15,18,24] The best results in the semi-automated peptide
synthesis were obtained by using 2-chlorotrityl resin loaded
with Fmoc-l-Ala-OH and an optimized microwave irradia-
tion protocol limiting the temperature to 408C. HATU and
HOAt were employed for coupling of the heavily glycosylat-
ed amino acid building blocks to safeguard an efficient cou-
pling. The corresponding aglycons were synthesized accord-
ing to a standard protocol with a maximum temperature of
788C.
Chem. Eur. J. 2012, 00, 0 – 0
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
3
&
ÞÞ
These are not the final page numbers!

N. Sewald et al.
charide derivative per azide) in
the presence of substoichiomet-
ric amounts of [CuACHTNUTRGNEUNG(MeCN)4]PF6
and TBTA (tris[(1-benzyl-1H-
1,2,3-triazol-4-yl)methyl]a-
mine)[27] under microwave irra-
diation at 608C for 2 h.[26] After
a few washing steps of the resin
with a mixture of sodium dieth-
yldithiocarbamate in DMF with
iPrNEt2, DMF and methylene
chloride, the peptides were
cleaved from the resin by
adding a mixture of TFA/H2O/
TIS (95:2.5:2.5). The peptides
were precipitated and dried
under reduced pressure, dis-
solved in water and lyophilized
to give white foams. The resi-
due was purified by preparative
HPLC to give pure AFGP ana-
logues
with
triazole-linked
mono- and disaccharide moiet-
ies (21–28).
CD experiments were con-
ducted to gain insight into the
conformational properties of
the synthetic AFGP analogues
and their corresponding agly-
cons in solution. The CD curves
of all Pro-containing peptides
show minima at about lꢁ195–
199 nm, but only for the glyco-
peptides (11, 13, 15) maxima at
about
lꢁ218–221 nm
(Figure 1). The typical features
in CD spectra of peptides with
PPII helical conformation are
the negative CD effect between
195 and 210 nm (depending on
the proline content) and the
slightly positive CD effect be-
tween 210 and 220 nm. All
spectra hint toward a threefold
left-handed helix[28] similar to a
Scheme 3. Synthetic AFGP analogues obtained containing monosaccharide-substituted Thr together with Pro
residues in different positions (11, 13, 15) and the corresponding aglycons (12, 14, 16).
polyproline
type
structure
(PPII)[29] or a conformational
The synthesis of the AFGP analogues with triazole-linked
mono- and disaccharide moieties (Scheme 5) was accom-
plished by adapting a previously published synthetic proce-
dure.[23b–d,26] The Fmoc-protected amino acids Pro and Azp
were successively coupled to the resin with HCTU and
iPr2NEt. The Fmoc-protecting group was cleaved after each
coupling step by treatment with piperidine in DMF.
mixture with random coil.[30] As expected, the tertiary amide
chromophores of Pro give rise to a red-shift of the minimum
compared to glycopeptides without any Pro residues in their
sequence.[15,18] The secondary amino group and the five-
membered ring of the Pro residues restrict the torsion
angles f to ꢀ658ꢂ158 supporting the formation of a PPII
helical structure.[29]
The solid phase bound aglycon peptide was treated with
the propargyl saccharides (2.5 equiv propargyl monosacchar-
ide derivative per azide function; 1.5 equiv propargyl disac-
The glycosylated peptides 11, 13 and 15 contain a differ-
ent number of Pro residues in different positions and exhibit
variable red-shifts of the CD minima. The AFGP analogue
&
4
&
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Modified AFGP Analogues
FULL PAPER
Scheme 5. Synthetic AFGP analogues with triazole-linked mono- and dis-
accharide moieties (21–28).
containing only one Pro residue (15) shows the least red-
shift with a negative band at lꢁ195 nm and a positive band
at lꢁ218 nm. The isomeric glycopeptides containing two
Pro residues in sequence positions 7 and 13 (11) or 10 and
13 (13) give rise to similar, but slightly changed CD curves;
this indicates that the position of the Pro residues has an in-
fluence on the conformation. Peptide 13 comprises a Pro
pair in positions i and i+3 and adopts a PPII helix. This
gives rise to CD curves with a shift to higher wavelength.[31]
Ala is also known to have a high propensity for a PPII
helix.[32] Interestingly, it was shown by ATR-FT-IR studies
that the a-helical content of AFGP increased at the ice/
water interface upon freezing of a supercooled solution film
at ꢀ158C.[33]
Scheme 4. Synthetic AFGP analogues obtained containing monosacchar-
ide-substituted Thr together with monosaccharide-substituted Hyp resi-
dues in different positions (17, 19) and the corresponding aglycons (18,
20).
Aglycons 12, 14, and 16 give rise to CD spectra with a
slightly more intense negative band in comparison to the
glycosylated counterparts, but lack a positive CD effect be-
tween 210 and 230 nm. Similar differences in CD between
glycosylated and unglycosylated peptides were already de-
scribed by Peltier et al.[17] Aglycons 12, 14, and 16 presuma-
bly adopt PPII helical conformation to a lesser extent; obvi-
ously, glycosylation stabilizes the PPII helix.[29,34] This is par-
ticularly true for the aglycon exclusively containing Ala-
Ala-Thr units, as reported earlier by us. The CD spectrum
of the peptide revealed a characteristic b-strand or b-turn
motif.[18] The carbohydrate moiety N-acetyl-d-galactosamine
seems to have an important impact on conformational sta-
bility due to additional interactions, such as hydrogen bonds
and the presence of water pockets/bridges established be-
tween the monosaccharide and the peptide backbone.[35] In-
terestingly, no difference in the position of the negative
band, that is, no red-shift is observed among the aglycons
Figure 1. CD spectra of the monosaccharide-substituted Pro-containing
AFGP analogues (11, 13, 15) and their corresponding aglycons (12, 14,
16) at a concentration of 0.2 mgmLꢀ1 in water at 208C (m=GalNAc).
Chem. Eur. J. 2012, 00, 0 – 0
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
5
&
ÞÞ
These are not the final page numbers!

N. Sewald et al.
with increasing Pro content (12, 14, 16). All three curves
show a negative CD effect at about lꢁ196–197 nm.
residues, but differ in the carbohydrate moiety at Thr
(AFGP8: b-d-galactosyl-(1!3)-a-N-acetyl-d-galactosamine;
11: a-N-acetyl-d-galactosamine). Although both peptides
are derivatized with different carbohydrate moieties, both of
them give rise to nearly identical CD spectra. The additional
carbohydrate moieties do not seem to further influence the
conformation according to the CD spectra.
The influence of regular incorporation of proline residues
on AFGP conformation was further probed by comparison
of the CD spectra of the prototype monosaccharide-substi-
tuted AFGP [AATm]4AA (29) with the Pro-containing gly-
copeptide [PATm]4PA (30).[18] The former clearly shows the
prevalence of a PPII helix with the highest magnitude in the
minimum and maximum, whereas the CD spectrum of the
latter exhibits a red-shifted, less intense minimum and does
not show a maximum at all. The high Pro content dislocates
the CD curve to a higher wavelength due to the tertiary
amide chromophore.
Fibrous collagen with the repetitive sequence (Gly-Pro-
Hyp) represents another prominent protein known to adopt
PPII helical conformation.[30] Therefore, collagen–AFGP
hybrid peptides 17 and 19 glycosylated with a-N-acetyl-d-
galactosamine at the hydroxyproline residues were synthe-
sized. Glycopeptide 17 exhibits a negative CD effect at l
ꢁ199 nm. The minimum is significantly more red-shifted
than that of peptide 19 with a negative band at about l
ꢁ195 nm (Figure 3). The position of the negative CD effect
of 19 is comparable with glycopeptide 15, which has one Pro
residue. The CD spectra of the corresponding aglycons 18
and 20 are similar to the aglycons 12, 14, and 16. They show
a more intense negative CD than the glycosylated peptides,
17 and 19, but a less intense positive band at lꢁ220 nm, in-
dicating a PPII helix albeit with lower propensity than the
glycopeptides (Figure 3). Nevertheless, both aglycons exhibit
a PPII-like conformation, regardless of the number of the
Gly-Pro-Hyp units, with a minimum at about lꢁ198 nm
and a marginal visible red-shift in comparison to each other.
A CD signature with a minimum closely below lꢁ200 nm
and a maximum at approximately 218 nm is usually associat-
ed with random-coil structures.[30] Temperature-dependent
CD experiments were accomplished in water between 08C
and +808C to further probe the existence of a PPII helix.
Tiffany and Krimm proposed a conformational connection
between a PPII helix and a random coil. In temperature-de-
pendent CD experiments of polypeptides and collagen at
least locally ordered extended helix structure was found.[36]
This was confirmed by experimental data by Dukor and
Keiderling, who compared vibrational CD with electronic
CD of polyproline peptides. It was concluded that the
random coil structure has a large fraction of helical regions
similar to the PPII conformation.[37] The CD spectra record-
ed at different temperatures (Figures S16–S21 in the Sup-
porting Information) exhibit one isodichroic point at about
lꢁ207–208 nm for both the Pro-containing glycopeptides
(11, 13, 15) and the corresponding aglycons (12, 14, 16). This
hints towards an equilibrium between at least two conform-
ers, presumably between a PPII helix and a random
coil.[15,18] CD difference spectra (e.g., Figure S34 in the Sup-
porting Information) in which the spectrum recorded at the
lowest temperature (08C) is subtracted from the spectrum
recorded at the highest temperature (+808C) might be indi-
cative of b-structure-like motifs being present at higher tem-
perature.[38] The PPII helix and b-strand conformers are very
close in the alanine dipeptide f/y energy map.[39] These re-
sults can be interpreted with a change in conformation from
a PPII helix to an assumedly b-turn-like-structure, accompa-
nied by random-coil contributions.[15,38a,40]
Comparison of the synthetic Pro-containing peptide 11
with the highly similar AFGP8, isolated from natural sour-
ces illustrates the conformational similarity according to CD
(Figure 2).[15] AFGP8 and 11 share the number of amino
acids and the sequence, especially with respect to the Pro
Figure 2. CD spectra of the monosaccharide-substituted Pro-containing
Figure 3. CD spectra of the monosaccharide-substituted peptides consist-
AFGP analogue (11), the corresponding natural AFGP8 and AFGP ana-
ing of Gly-Pro-HypACTHNGUTERN(UNG GalNAc) tripeptide units (17, 19) and their corre-
logues [AAT
G
N
sponding aglycons (18, 20) with a concentration of 0.2 mgmLꢀ1 in water
at 208C (m=GalNAc).
at 208C (m=GalNAc; d=Gal-GalNAc).
&
6
&
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Modified AFGP Analogues
FULL PAPER
As mentioned before for the Pro-containing peptides, the
CD of the glycopeptides 17 and 19 shows a significant differ-
ence in the position of the negative band, whereas the CD
curves of the corresponding aglycons 18 and 20 are only
slightly changed.
residues gave a slightly different CD spectrum with a lower
intensity of the negative band but a higher one of the posi-
tive band, accompanied by a distinct red-shift (Figure 4).
The related oligoprolines 25–28 carrying the disaccharide
b-d-galactosyl-(1!3)-a-N-acetyl-d-galactosamine connected
across a triazole moiety at every third residue, give rise to
CD spectra very similar to 21–24. The PPII helical confor-
mation is indicated by a minimum at lꢁ203–206 nm and a
positive band at lꢁ223–227 nm (Figure 5). The additional
The temperature-dependent CD experiments recorded in
the range from 08C to 808C (Figures S22–S25 in the Sup-
porting Information) also show isodichroic points between
205 and 209 nm for the collagen–AFGP hybrid peptides 17–
20, indicating an equilibrium between two conformations.
Again, the aglycons are characterized by higher propensity
for random coil than the glycopeptides. The difference spec-
tra obtained by subtracting the CD spectrum recorded at
08C from the one recorded at 808C indicate a b-turn or b-
structure-like motif. The Gly-Pro-Hyp containing peptides
adopt a PPII structure in a mixture with random coil at low
temperatures. With increasing temperature the overall struc-
ture shows a higher contribution of unordered conformation
accompanied by b-turn-like conformations.[15,38a,39]
The oligoprolines 21–24 containing the a-N-acetyl-d-gal-
actosamine motifs at every third residue, connected across a
triazole moiety, represent a further modification. The CD
spectra show the presence of a PPII helical structure, albeit
with distinct differences regarding the intensity of the
minima and maxima. The molar ellipticity per residue in-
creases as expected with the number of amino acid residues.
Moreover, the configuration of the triazolylproline (Tzp)
also has an influence. Peptides 21, 22 and 23 all contain
2S,4S configured Tzp residues. Peptides 21 and 22 reveal a
negative band at about lꢁ204 and 205 nm and positive
maxima at lꢁ223 and 224 nm, respectively. Peptide 23 with
the longest chain length shows the highest intensity in nega-
tive CD effect and in the red-shift with a negative band at l
ꢁ206 nm and a positive band at lꢁ224 nm (Figure 4).[41]
The intensity of the negative band and the red-shift increase
with the number of repeating units in the peptide; this is in
accordance with an increasing stabilization of the PPII
helix.[42] Peptide 24 with 2S,4R configured triazolylproline
Figure 5. CD spectra of the disaccharide-substituted oligoprolines (25–28)
at a concentration of 0.2 mgmLꢀ1 in water at 208C (d=Gal-GalNAc).
b-d-galactosyl moiety does not seem to have significant in-
fluence on the overall peptide conformation as it was ob-
served for the previously discussed AFGP. In almost the
same manner, the minima and maxima are red-shifted and
the intensity of the minimum increases with the length of
the glycopeptide 25–27. The 2S,4R configured triazolylpro-
line containing peptide 28, like analogue 24, gives rise to a
CD spectrum conformation with the highest red-shift in
comparison to the 2S,4S configured peptides but with the
lowest intensity of the minimum at lꢁ206 nm.
Temperature-dependent CD spectra in the temperature
range from 08C to +808C were recorded to obtain further
insight into the conformational features of the AFGP ana-
logues (Figures S26–S33 in the Supporting Information).
The CD spectra accumulated for all peptides of this type
(21–28) exhibit two isodichroic points at lꢁ192–193 and
212–214 nm, respectively. The isodichroic point at shorter
wavelength often is not reported because of problems with
the absorbance of the sample below 195 nm. The CD spec-
tra decrease in intensity with higher temperature, as men-
tioned above, for the Pro-containing AFGP ana-
logues.[15,36,37] However, the pronounced CD effects suggest
a higher stability of the PPII helix than was observed for the
Pro-containing peptides (11–16).[43] The intensity changes
with temperature by only about 20%. Especially the pepti-
des with the 2S,4R configured triazolylproline residues (24,
28) exhibit somewhat lower temperature dependence; this is
indicative of a higher conformational stability. This could
Figure 4. CD spectra of the monosaccharide-substituted oligoprolines
(21–24) at
GalNAc).
a
concentration of 0.2 mgmLꢀ1 in water at 208C (m=
Chem. Eur. J. 2012, 00, 0 – 0
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
7
&
ÞÞ
These are not the final page numbers!

N. Sewald et al.
arise from the higher preference of (2S,4R)-Tzp compared
to (2S,4S)-Tzp for a trans amide bond conformation which is
required for the PPII conformation.[26a]
In summary, the Pro-containing peptides 11–16 and the
collagen motif-containing peptides (17–20), upon heating
undergo a structural transition from a mixture of a PPII and
random coil potentially to a mixture of a b-structure and
random coil, whereas the oligoprolines carrying either a
mono- or a disaccharide (21–28) basically retain their PPII
conformation upon heating. In the first two types of pepti-
des the PPII helical structure is presumably disrupted with
short fractions of unordered peptide at low temperature,
whereas at high temperature the conformational mixture
consist of b-structures with short random coil structures.[36a]
The most stable PPII conformation is found in the oligopro-
lines (21–28) as indicated by the more intensive bands. This
higher stability most probably arises by the higher proline
content of these peptides and the amidated C terminus.[44] In
turn, the AFGP analogues 11, 13, 15, 17, and 19 adopt a
PPII helix to a larger extent in comparison to their agylcons;
this is probably explicable with additionally formed interac-
tions by the N-acetyl group of the carbohydrate moieties,
but also by the formation of water pockets/bridges between
the carbohydrate and peptide moieties.[15,36,39c] More Pro res-
idues present in the sequence do most likely favor a PPII
helical conformation at the expense of an unordered struc-
ture;[30] this is mostly apparent in the slight red-shift of the
CD curves on the one hand and in the high magnitude in
positive CD effect in the higher wavelength region at about
lꢁ220 nm.[41c]
The activity and efficiency of antifreeze agents can be
tested in ice recrystallization experiments.[3c,45–47] The anti-
freeze activity of the synthesized peptides was investigated
in a previously described ice recrystallization inhibition
assay, which allows for the separation of the antifreeze
effect from other factors that influence recrystallization,
such as the ice volume fraction.[47] In this assay all peptides
are dissolved in 45 wt% sucrose solutions and the recrystal-
lization kinetics is determined at ꢀ88C. The recrystallization
rate constant, k, is determined according to the Lifshitz–
Slyozov–Wagner (LSW) theory by analyzing the cubic mean
crystal radius, r3, as a function of the recrystallization time,
Figure 6. a) Recrystallization rate constant, kl0, as a function of the pep-
tide concentration, c. Shown are the mean values of typically 3–5 individ-
ual experiments represented by symbols. The corresponding standard
errors are indicated by error bars. Active peptides show a decrease of the
rate constant, kl0, down to negligible values. When kl0 becomes half of
the diffusion-limited value of pure sucrose solutions the corresponding
concentration is defined at the inhibitor concentration, ci, due to the ice-
binding ability of antifreeze agents. b) Inhibitory efficiency represented
by the inhibitor concentration, ci. For peptides that were inactive in the
tested concentration range arrows indicate the highest investigated con-
centrations.
3
t: r3 =r0 +kt, where r0 is the cubic mean radius at time t=
0 min. Several experimental parameters that influence the
rate constant, k, can be neglected when all experiments are
performed with the same basic solute (sucrose), at the same
temperature (ꢀ88C), and at small ice volume fractions. In
that case the observed rate constants in each experiment all
converge to the theoretical diffusion-limited LSW rate con-
stant at zero volume fractions in sucrose solutions at ꢀ88C.
Diffusion-limited implies that the diffusion of water mole-
cules between the ice crystals is the slowest and, hence, rate-
limiting process that is responsible for the observed ice re-
crystallization rate. We also take into account the effects of
the nonzero ice volume fraction.[46] The resulting rate con-
stant, kl0, is then dependent upon the presence, the efficien-
cy, and the concentration of ice-binding agents (Figure 6a).
When the ice surface coverage of ice-binding agents be-
comes sufficiently large, the limiting process for recrystalli-
zation changes from water diffusion to liquid-to-ice transfer.
This changeover can be used to determine the peptide-spe-
cific inhibition concentration, ci, which we define as that
concentration at which kl0 has decreased to half of the diffu-
&
8
&
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Modified AFGP Analogues
FULL PAPER
sion-limited value. This definition of ci is illustrated in Fig-
ure 6a in the case of [AATm]4AA (29).
peptoids that contain similar groups did not show any activi-
ty either.[16] However, it has to be mentioned that there are
several structural differences between these analogues and
the natural AFGP. Hence, at present it is not possible to
verify this hypothesis, but we can conclude that a periodic
amino acid sequence and a PPII helical structure are not cri-
teria that are sufficient for antifreeze activity. The CD
curves of the oligoprolines 21–28 and the collagen analogues
17 and 19 all indicate the existence of a PPII helix. Never-
theless, all these peptides do not inhibit ice recrystallization
up to concentrations of 1000 mgmLꢀ1. In the case of the
Gly-Pro-Hyp containing peptides the main reason might be
the substitution of Thr by Hyp. It was shown by Tachibana
et al. that removal of the Thr g-methyl group (substitution
of Thr by Ser) suffices to lead to a significant reduction in
antifreeze activity.[11] It could also be possible that the
amino acid substitution contributes to a reduced recrystalli-
zation inhibition efficiency, and that the sum of these reduc-
tions leads to a non-efficient agent.
Figure 6b shows the resulting ci values of selected pepti-
des from this study, and also for comparison, those of previ-
ously investigated peptides.[18] At a ci value of 20 mmolLꢀ1
(41 mgmLꢀ1) the peptide [AATm]4AA (29) without any Pro
is the most efficient ice recrystallization inhibitor within this
series. The introduction of Pro for Ala does lead to a de-
crease in the inhibition efficiency regardless of the position
or the number of exchanges, even though the antifreeze ac-
tivity is retained. However, the Gly-Pro-Hyp containing
peptides (17, 19) and the polyproline peptides (23, 27, 28)
do not show any antifreeze activity in the investigated con-
centration range.
Our previous investigations did not reveal any significant
difference in the efficiency of the natural AFGP8
([AATd]2PATdAATdPA), and its synthetic analogue
[AATd]4AA, which does not contain any Pro.[15] The associ-
ated ci values are 0.086 mmolLꢀ1 (0.22 mgmLꢀ1) and
0.085 mmolLꢀ1 (0.23 mgmLꢀ1), respectively. In contrast, the
ci values of the corresponding monosaccharide AFGP ana-
logues are 190 mmolLꢀ1 (400 mgmLꢀ1) for the Pro-containing
peptide 11 and 20 mmolLꢀ1 (41 mgmLꢀ1) for [AATm]4AA
(29). Thus, in this case the introduction of Pro at positions 7
and 13 leads to a recrystallization inhibitor that is less effi-
cient by about one order of magnitude. These comparisons
imply that the less efficient monosaccharide AFGP ana-
logues are much more sensitive to backbone mutations than
the natural disaccharide AFGP, that is, they reveal stronger
effects on the recrystallization inhibition efficiencies. Similar
structural differences are not observable as the addition of
b-d-galactosyl moieties does not alter the CD spectra signifi-
cantly. The change from [AATm]4AA to the [Pro7,Pro13] de-
rivative is similar to that from [AATd]4AA to AFGP8.
Interestingly, peptide 11 is the antifreeze agent with the
lowest efficiency in this series. The ci value of peptide 13 is
49 mmolLꢀ1 (100 mgmLꢀ1). This peptide contains the same
amount of Pro as peptide 11, but at positions 10 and 13.
Peptide 15, which contains Pro only at position 13, is not
more efficient than peptides 11 and 13, which contain two
Pro residues. The ci value is 130 mmolLꢀ1 (280 mgmLꢀ1). It
seems that Pro introduction is less disturbing if it is posi-
tioned in neighboring tripeptide units. With 81 mmolLꢀ1
(ci =180 mgmLꢀ1), [PATm]4PA (30) is less efficient than the
[Pro10,Pro13] derivative 13, but both are more efficient than
peptides 11 and 15. This result is in agreement with our pre-
vious statement that irregular substitutions of Pro for Ala
may prevent the formation of a periodic, active conforma-
tion.[18] This may be due to non-periodic turn formations
around the Pro residues.
Conclusion
Monosaccharide- and disaccharide-substituted AFGP ana-
logues were synthesized that contained amino acid substitu-
tions (Ala!Pro). The influence of the saccharide portion
(mono- vs. disaccharide) and of the Thr!Hyp exchange was
investigated. Furthermore, glycosylated (Pro-Tzp-Pro)n-pep-
tides were assembled giving triazole-linked oligoprolines fur-
nished either with a mono- or disaccharide. The conforma-
tion of these glycopeptides was examined by using CD and
compared with their corresponding aglycons. In microphysi-
cal antifreeze assays the activity of all glycopeptides was
compared.
Direct comparison of the monosaccharide- with the disac-
charide-substituted AFGP analogues did not reveal any
change in conformation but in antifreeze activity.[15,18] The
replacement of Ala by Pro residues in the peptide backbone
of AFGP8 did not significantly lower the activity relative to
Pro-free AFGP. In contrast, the incorporation of Pro in
monosaccharide-substituted AFGP analogues led to de-
creased antifreeze effectiveness. The carbohydrate moiety
overrides the influence of Pro residues present in the se-
quence. The influence of the disaccharide moiety in AFGP8
on antifreeze activity is so pronounced that the mutation in
the sequence does not affect the ability to inhibit ice crystal
growth. The monosaccharide-substituted AFGP analogues
with Pro mutations display lower activity than Pro-free gly-
copeptides.
Although the collagen motif-containing peptides (17, 19)
assumably adopt PPII helical conformation with similar pro-
pensity as the Pro-containing AFGP analogues (11, 13, 15),
the ability to inhibit ice crystal growth was found only for
the latter. High stability in PPII helical conformation obvi-
ously is not the only precondition for an effective antifreeze
agent. The oligoprolines 21–28 lack antifreeze activity, pre-
sumably because of the larger triazole spacer between the
In contrast to the hitherto discussed Pro-containing
AFGP analogues the oligoprolines 23, 27, and 28 are totally
inactive in the investigated concentration range. The highest
concentrations that were tested are in the range of
400 mmolLꢀ1 (1000 mgmLꢀ1). The observed inactivity may
result from the linkage between the backbone and the car-
bohydrate moieties, the triazole unit. Previously investigated
Chem. Eur. J. 2012, 00, 0 – 0
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
9
&
ÞÞ
These are not the final page numbers!

N. Sewald et al.
[7] a) J. Garner, M. M. Harding, ChemBioChem 2010, 11, 2489–2498;
b) R. E. Feeney, Am. Sci. 1974, 62, 712–719.
[8] a) Y. Lin, J. G. Duman, A. L. DeVries, Biochem. Biophys. Res.
Commun. 1972, 46, 87–92; b) H. R. Morris, M. R. Thompson, D. T.
Osuga, A. I. Ahmed, S. M. Chan, J. R. Vandenheede, R. E. Feeney,
J. Biol. Chem. 1978, 253, 5155–5162.
[9] T. S. Burcham, D. T. Osuga, B. N. Rao, C. A. Bush, R. E. Feeney, J.
Biol. Chem. 1986, 261, 6384–6389.
[10] A. P. A. Wçhrmann, Mar. Ecol. Prog. Ser. 1996, 130, 47–59.
[11] Y. Tachibana, G. L. Fletcher, N. Fujitani, S. Tsuda, K. Monde, S.-I.
Nishimura, Angew. Chem. 2004, 116, 874–880; Angew. Chem. Int.
Ed. 2004, 43, 856–862.
[12] T. Tsuda, S.-I. Nishimura, Chem. Commun. 1996, 2779–2780.
[13] Y. Tachibana, N. Matsubara, F. Nakajima, S. Tsuda, K. Monde, S.-I.
Nishimura, Tetrahedron 2002, 58, 10213–10224.
[14] P. H. Tseng, W. T. Jiaang, M. Y. Chang, S. T. Che, Chem. Eur. J.
2001, 7, 585–590.
[15] L. Nagel, C. Plattner, C. Budke, Z. Majer, A. L. DeVries, T. Berke-
meier, T. Koop, N. Sewald, Amino Acids 2011, 41, 719–732.
[16] A. S. Norgren, C. Budke, Z. Majer, C. Heggemann, T. Koop, N.
Sewald, Synthesis 2009, 3, 488–494.
[17] R. Peltier, C. W. Evans, A. L. DeVries, M. A. Brimble, A. J. Dingley,
D. E. Williams, Cryst. Growth Des. 2010, 10, 5066–5077.
[18] C. Heggemann, C. Budke, B. Schomburg, Z. Majer, M. Wißbrock, T.
Koop, N. Sewald, Amino Acids 2010, 38, 213–222.
carbohydrate moiety and peptide backbone. AFGP seem to
be able to structure the surrounding water in the first hydra-
tion shell[36b] and to perturb the hydration dynamics.[6] The
carbohydrate residues in AFGP are assumed to be aligned
on one side of the PPII helix. However, this alignment does
not suffice for antifreeze activity, as shown here for exam-
ple, for Gal-GalNAc derivatized PPII helical analogues.
Experimental Section
Experimental procedures and the full characterization of all intermedi-
ates and the final products are described in the Supporting Information.
Microphysical ice recrystallization analysis: The antifreeze activity was
determined based on a previously described method in which the inhibi-
tory effect of antifreezes on ice recrystallization is quantified. Briefly, the
peptides (TFA salts) were dissolved in sucrose (45 wt%) solution at vari-
ous concentrations; aliquots of these solutions (2 mL) were placed be-
tween two glass cover slides with a film thickness of about 10–20 mm, and
the resulting samples were positioned on a temperature-controlled silver
block inside a cold stage (Linkam MDBCS 196) mounted on an optical
microscope (Olympus BX51) in bright field transmission mode. The sam-
ples were cooled to ꢀ508C at a rate of 208Cminꢀ1, reheated to ꢀ88C at
a rate of 108Cminꢀ1, and annealed at this temperature for 2 h. The re-
crystallization of the resulting polycrystalline ice was recorded by taking
images with a digital video camera in 12 s time intervals while the images
were analyzed simultaneously by using a LabVIEW virtual instrument.
The ice recrystallization rate was determined by using LSW theory and
then used to obtain the inhibitor-related concentration, ci, at which the
recrystallization limiting process turns over from water diffusion to
liquid-to-ice transfer due to the presence of ice-binding antifreezes.[46]
[19] N. Miller, G. M. Williams, M. A. Brimble, Org. Lett. 2009, 11, 2409–
2412.
[20] C. A. Bush, S. Ralapati, G. M. Matson, R. B. Yamasaki, D. T. Osuga,
Y. Yeh, R. E. Feeney, Arch. Biochem. Biophys. 1984, 232, 624–631.
[21] C. A. Bush, R. E. Feeney, Int. J. Pept. Protein Res. 2009, 28, 386–
397.
[22] a) A. N. Lane, L. M. Hays, L. M. Crowe, J. H. Crowe, R. E. Feeney,
Protein Sci. 1998, 7, 1555–1563; b) A. N. Lane, L. M. Hays, N.
Tsvetkova, R. E. Feeney, L. M. Crowe, J. H. Crowe, Biophys. J.
2000, 78, 3195–3207.
[23] a) R. S. Erdmann, H. Wennemers, Angew. Chem. 2011, 123, 6967–
6970; Angew. Chem. Int. Ed. 2011, 50, 6835–6838; b) Y. A. Nagel,
M. Kuemin, H. Wennemers, Chimia 2011, 65, 264–267; c) R. S. Erd-
mann, M. Kumin, H. Wennemers, Chimia 2009, 63, 197–200; d) M.
Kꢁmin, L. S. Sonntag, H. Wennemers, J. Am. Chem. Soc. 2007, 129,
466–467; e) M. Kuemin, Y. A. Nagel, S. Schweizer, F. W. Monnard,
C. Ochsenfeld, H. Wennemers, Angew. Chem. 2010, 122, 6468–6471;
Angew. Chem. Int. Ed. 2010, 49, 6324–6327; f) M. Kuemin, J. Engel,
H. Wennemers, J. Pept. Sci. 2010, 16, 596–600.
Acknowledgements
The authors gratefully acknowledge support from Deutsche Forschungs-
gemeinschaft (SFB 613, Project A8). L.N. was funded by a PhD fellow-
ship of Bielefeld University. Zsuzsa Majer (Eçtvçs Lorꢃnd University
Budapest) contributed helpful suggestions.
[24] C. Plattner, M. Hçfener, N. Sewald, Org. Lett. 2011, 13, 545–547.
[25] S. Hanessian, P. Lavallee, Can. J. Chem. 1975, 53, 2975–2977.
[26] a) R. S. Erdmann, H. Wennemers, Org. Biomol. Chem. 2012, 10,
1982–1986; b) R. S. Erdmann, H. Wennemers, J. Am. Chem. Soc.
2010, 132, 13957–13959.
[1] a) P. L. Davies, B. D. Sykes, Curr. Opin. Struct. Biol. 1997, 7, 828–
834; b) M. M. Harding, P. I. Anderberg, A. D. Haymet, Eur. J. Bio-
chem. 2003, 270, 1381–1392.
[27] a) Q. Wang, T. R. Chan, R. Hilgraf, V. V. Fokin, K. B. Sharpless,
M. G. Finn, J. Am. Chem. Soc. 2003, 125, 3192–3193; b) C. W.
Tornøe, C. Christensen, M. Meldal, J. Org. Chem. 2002, 67, 3057–
3064; c) V. V. Rostovtsev, L. G. Green, V. V. Fokin, K. B. Sharpless,
Angew. Chem. 2002, 114, 2708–2711; Angew. Chem. Int. Ed. 2002,
41, 2596–2598.
[28] C. A. Bush, R. E. Feeney, D. T. Osuga, S. Ralapati, Y. Yeh, Int. J.
Pept. Protein Res. 2009, 17, 125–129.
[29] J. Makowska, S. Rodziewicz-Motowidlo, K. Baginska, J. A. Vila, A.
Liwo, L. Chmurzynski, H. A. Scheraga, Proc. Natl. Acad. Sci. USA
2006, 103, 1744–1749.
[2] J. Barrett, Int. J. Biochem. Cell Biol. 2001, 33, 105–117.
[3] a) K. V. Ewart, Q. Lin, C. L. Hew, Cell. Mol. Life Sci. 1999, 55, 271–
281; b) Y. Yeh, R. E. Feeney, Chem. Rev. 1996, 96, 601–618; c) C. A.
Knight, J. Hallet, A. L. DeVries, Cryobiology 1988, 25, 55–66; d) A.
Parody-Morreale, K. P. Murphy, E. Dicera, R. Fall, A. L. DeVries,
S. J. Gill, Nature 1988, 333, 782–783; e) J. A. Raymond, P. Wilson,
A. L. DeVries, Proc. Natl. Acad. Sci. USA 1989, 86, 881–885.
[4] a) C. A. Knight, A. L. DeVries, L. D. Oolman, Nature 1984, 308,
295–296; b) V. Bouvet, R. N. Ben, Cell Biochem. Biophys. 2003, 39,
133–144; c) R. Peltier, M. A. Brimble, J. M. Wojnar, D. E. Williams,
C. W. Evans, A. L. DeVries, Chem. Sci. 2010, 1, 538–551.
[30] A. Rath, A. R. Davidson, C. M. Deber, Biopolymers 2005, 80, 179–
185.
[31] B. J. Stapley, T. P. Creamer, Protein Sci. 2008, 8, 587–595.
[32] A. L. Rucker, C. T. Pager, M. N. Campbell, J. E. Qualls, T. P. Cream-
er, Proteins 2003, 53, 68–75.
[5] a) C. A. Knight, C. C. Cheng, A. L. DeVries, Biophys. J. 1991, 59,
409–418; b) A. D. J. Haymet, L. G. Ward, M. M. Harding, J. Am.
Chem. Soc. 1999, 121, 941–948; c) N. Holmberg, G. Lilius, L. Bꢁlow,
FEBS Lett. 1994, 349, 354–358; d) B. P. Duncker, S. Y. Gauthier,
P. L. Davies, Biochem. Biophys. Res. Commun. 1994, 203, 1851–
1857; e) M. C. Loewen, X. Liu, P. L. Davies, A. J. Dauglius, Appl.
Microbiol. Biotechnol. 1997, 48, 480–486.
[33] Y. Uda, S. Zepada, F. Kaneko, Y. Matsuura, Y. Furukawa, J. Phys.
Chem. B 2007, 111, 14355–14361.
[6] S. Ebbinghaus, K. Meister, B. Born, A. L. DeVries, M. Gruebele, M.
Havenith, J. Am. Chem. Soc. 2010, 132, 12210–12211.
[34] a) A. F. Drake, G. Siligardi, W. A. Gibbsons, Biophys. Chem. 1988,
31, 143–146; b) M. G. Venugopal, J. A. Ramshaw, E. Braswell, D.
&
10
&
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Modified AFGP Analogues
FULL PAPER
Zhu, B. Brodsky, Biochemistry 1994, 33, 7948–7956; c) J. G. Bann,
H. P. Bachinger, D. H. Peyton, Biochemistry 2003, 42, 4042–4048;
d) J. G. Bann, H. P. Bachinger, J. Biol. Chem. 2000, 275, 24466–
24469.
Liu, K. Chen, A. Ng, Z. Shi, R. W. Woody, N. R. Kallenbach, J. Am.
Chem. Soc. 2004, 126, 15141–15150.
[39] H. J. Dyson, P. E. Wright, Annu. Rev. Biophys. Chem. 1991, 20, 519–
538.
[35] a) D. H. Nguyen, M. E. Colvin, Y. Yeh, R. E. Feeney, W. H. Fink,
Biophys. J. 2002, 82, 2892–2905; b) F. Corzana, J. H. Busto, G. Jime-
nez-Oses, J. L. Asensio, J. Jimenez-Barbero, J. M. Peregrina, A. Ave-
noza, J. Am. Chem. Soc. 2006, 128, 14640–14648; c) F. Corzana,
J. H. Busto, M. G. de Luis, A. Fernandez-Tejada, F. Rodriguez, J. Ji-
menez-Barbero, A. Avenoza, J. M. Peregrina, Eur. J. Org. Chem.
2010, 3525–3532; d) F. Corzana, J. H. Busto, G. Jimenez-Oses, M. G.
de Luis, J. L. Asensio, J. Jimenez-Barbero, J. M. Peregrina, A. Ave-
noza, J. Am. Chem. Soc. 2007, 129, 9458–9467; e) R. Y. Tam, C. N.
Rowley, I. Petrov, T. Zhang, N. A. Afagh, T. K. Woo, R. N. Ben, J.
Am. Chem. Soc. 2009, 131, 15745–15753; f) S. S. Mallajosyula, A. D.
MacKerell, Jr., J. Phys. Chem. B 2011, 115, 11215–11229.
[40] a) N. W. Owens, J. Stetefeld, E. Lattova, F. Schweizer, J. Am. Chem.
Soc. 2010, 132, 5036–5042; b) A. S. Tatham, A. F. Drake, P. R.
Shewry, Biochem. J. 1989, 259, 471–476.
[41] a) B. Brodsky, G. Thiagarajan, B. Madhan, K. Kar, Biopolymers
2008, 89, 345–353; b) J. Engel, H. P. Bꢀchinger, Top. Curr. Chem.
2005, 247, 7–33; c) G. B. Fields, D. J. Prockop, Biopolymers 1996, 40,
345–357.
[42] M. Rothe, H. Rott, Angew. Chem. 1976, 88, 844–845; Angew. Chem.
Int. Ed. Engl. 1976, 15, 770–771.
[43] J.-C. Horng, R. T. Raines, Protein Sci. 2006, 15, 74–83.
[44] M. Kuemin, S. Schweizer, C. Ochsenfeld, H. Wennemers, J. Am.
Chem. Soc. 2009, 131, 15474–15482.
[36] a) M. L. Tiffany, S. Krimm, Biopolymers 1968, 6, 1379–1382;
b) M. L. Tiffany, S. Krimm, Biopolymers 1969, 8, 347–359.
[37] R. K. Dukor, T. A. Keiderling, Biopolymers 1991, 31, 1747–1761.
[38] a) R. W. Woody, in The Peptides, Vol. 7 (Ed.: V. J. Hruby), Academic
Press, New York, 1985, pp. 15–114; b) K. Chen, Z. Liu, N. R. Kal-
lenbach, Proc. Natl. Acad. Sci. USA 2004, 101, 15352–15357; c) Z.
[45] C. A. Knight, D. Wen, R. A. Laursen, Cryobiology 1995, 32, 23–34.
[46] C. Budke, T. Koop, ChemPhysChem 2006, 7, 2601–2606.
[47] C. Budke, C. Heggemann, M. Koch, N. Sewald, T. Koop, J. Phys.
Chem. B 2009, 113, 2865–2873.
Received: June 14, 2012
Published online: && &&, 0000
Chem. Eur. J. 2012, 00, 0 – 0
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
11
&
ÞÞ
These are not the final page numbers!

N. Sewald et al.
Out of the cold: The antifreeze activity
of AFGP (antifreeze glycoprotein, see
figure) analogues strongly depends on
the amino acid composition. Synthetic,
near-native compounds (e.g., with pro-
line replacing alanine) retain the
effect, whereas others exhibit reduced
or abolished influence on ice recrystal-
lization, despite having high propensity
to adopt polyproline II helical confor-
mation, as proposed for the native
AFGP.
Glycopeptides
L. Nagel, C. Budke, R. S. Erdmann,
A. Dreyer, H. Wennemers, T. Koop,
N. Sewald* .................... &&&& — &&&&
Influence of Sequential Modifications
and Carbohydrate Variations in
Synthetic AFGP Analogues on
Conformation and Antifreeze Activity
&
12
&
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!