A Comprehensive Landscape for Fibril Association Behaviors Encoded Synergistically by Saccharides and Peptides

Article abstract of DOI:10.1021/jacs.1c01951

Nature provides us a panorama of fibrils with tremendous structural polymorphism from molecular building blocks to hierarchical association behaviors. Despite recent achievements in creating artificial systems with individual building blocks through self-

Full text of DOI:10.1021/jacs.1c01951

pubs.acs.org/JACS
Article
A Comprehensive Landscape for Fibril Association Behaviors
Encoded Synergistically by Saccharides and Peptides
Rongying Liu, Ran Zhang, Long Li, Zdravko Kochovski, Lintong Yao, Mu-Ping Nieh, Yan Lu,
Tongfei Shi, and Guosong Chen*
Cite This: J. Am. Chem. Soc. 2021, 143, 6622−6633
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Nature provides us a panorama of fibrils with tremendous structural polymorphism from molecular building blocks to
hierarchical association behaviors. Despite recent achievements in creating artificial systems with individual building blocks through
self-assembly, molecularly encoding the relationship from model building blocks to fibril association, resulting in controlled
macroscopic properties, has remained an elusive goal. In this paper, by employing a designed set of glycopeptide building blocks and
combining experimental and computational tools, we report a library of controlled fibril polymorphism with elucidation from
molecular packing to fibril association and the related macroscopic properties. The growth of the fibril either axially or radially with
right- or left-handed twisting is determined by the subtle trade-off of oligosaccharide and oligopeptide components. Meanwhile,
visible evidence for the association process of double-strand fibrils has been experimentally and theoretically proposed. Finally the
fibril polymorphs demonstrated significant different macroscopic properties on hydrogel formation and cellular migration control.
association behaviors starting from the molecular level of
building blocks has not yet been targeted. A comprehensive
landscape from molecular design to fibril structure and then
fibril association behaviors, resulting in controllable material
property, would greatly enhance our ability to design
functional fibrous materials.
Given the glycoprotein/proteoglycan nature of some fiber-
forming proteins such as collagen,15 α-synuclein,16 and
fibronectin,17 more importantly, both the saccharide moiety
and peptide part appear to be involved in fibril association of
glycoproteins with elusive explanations.15 For example, O-
GlcNAcylation was found to be able to alter the aggregation of
α-synuclein.18−20 It appears more rational to leverage
glycopeptide models to perform fibril association research
INTRODUCTION
■
Fibril associations are ubiquitous in key life processes, e.g.,
actin filaments cross-linking into actin bundles,1 collagen fibers
hierarchically associating into soft tissues,2 and amyloid fibrils
pathologically associating into amyloid plaque,3 to name a few.
The knowledge of the underlying mechanism of fibril
association is the foundation of understanding diverse key
life processes. Meanwhile, fibril associations also dictate
chirality,4 optical characteristics,5,6 and mechanical properties
of fibrous materials.7−9 Thus, the rationale of fibril association
is vitally important for the design and development of fibrous
materials.
Studies devoted to the fibril association have been emergent.
The phenomenon began to be understood by using statistical
analysis of atomic force images from the viewpoint of polymer
physics.10 Recently, some associated fibril species such as
bundles of intertwined fibers11 and braided fibrils12 have been
achieved based on some individual amphiphilic models that are
far from their natural blueprints in structural aspects. Despite
these cases having successfully provided a glimpse of fibril
association,13,14 a comprehensive picture of the fibril
Received: February 25, 2021
Published: April 26, 2021
https://doi.org/10.1021/jacs.1c01951
J. Am. Chem. Soc. 2021, 143, 6622−6633
© 2021 American Chemical Society
6622

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
than use simple short peptide models. Despite significant
achievements in creating artificial one-dimensional structures
through self-assembly of short peptides,21−24 the synergic
effect of saccharides and peptides within this landscape of fibril
association remains elusive.
sialyllactose (SL) by molecular simulation after considering the
complexity and availability of saccharides. It was observed that
the protofilament (starting structure) formed by the
glycopeptide model of SL-3F (Scheme 1 a, 1b) in a parallel
Here, we provide a family of glycopeptide models with
oligosaccharide and oligopeptide components. Both of the
components have a significant structural contribution to fibril
formation and the subsequent association. By controlling the
subtle balance of the hydrogen bonds between oligosaccharides
and those between oligopeptides as well as other interactions, a
library of typical fibril association behaviors has been exhibited,
such as single protofibrils laterally associated into flat
multistrand ribbons, laterally associated into twisted left-
handed double-strand fibrils, or radially associated into twisted
right-handed nanoribbons. By combining experimental evi-
dence and results from all-atom dynamics simulation, the
whole process from fibril formation to association and related
properties has been well elucidated by the competition
between strand-forming and strand-association, the competi-
tion between the hydrogen bonds from oligosaccharides or
oligopeptides, and a set of antagonistic interactions between
electrostatic repulsion and hydrophobic interaction. In
addition, visible evidence for the association process of
double-strand fibrils has been experimentally and theoretically
proposed. The links between fibril association behavior and
some macroscopic properties such as hydrogelation and cell−
matrix interaction have also been identified. In short, the
present work would shed light on breaking the molecular code
governing fibril association and would open up tremendous
opportunities for the design of fibrous materials.
Scheme 1. (a, b) Representative Model of SL-3F, All-Atom
Molecular Dynamics Simulation Results for (c) Antiparallel
Packing Mode and (d) Parallel Packing Mode, Quantified
Comparison of Hydrogen Bonds Formed by (e) Saccharide
Part (Saccharide H-Bonds), and (f) Two Components
(Saccharide Part and Peptide Moiety) (Total H-Bonds) in
Parallel Packing Mode and Antiparallel Packing Mode (12
× SL-3F)
RESULTS AND DISCUSSION
■
Selection and Fabrication of Models for Fibril
Association. To achieve a comprehensive landscape of fibril
association behaviors, a set of glycopeptide models with the
possibility of associating into different fibrils is required. In the
literature, short peptides have been employed as model
compounds to prepare various fibrils, in which the peptide
moieties either preferably adopt antiparallel packing man-
ner25,26 or occasionally adopt a parallel packing manner.27 Due
to the lack of another component that can reasonably check
and balance peptide component, these unbalanced peptidic
models with the propensity to adopt antiparallel packing
manner seems to be difficult to map a comprehensive
landscape of fibril association behaviors. To some extent, it
could be that although fibrils based on short peptides have
been widely investigated, the related fibril association behavior
has not been understood at the molecular level.
To solve this problem, we sought to construct a balanced
glycopeptide model with a propensity to adopt a parallel
packing manner, instead of antiparallel packing, in the fibril of
the glycopeptide. For this type of balanced glycopeptide
model, the interactions between saccharides should be
compatible with the interactions between peptides. Meanwhile,
the spatial size of the two parts should be compatible too. To
access this type of glycopeptide model, triphenylalanine
(denoted as 3F) was first selected as the peptide moiety
concerning its proven potential for fibril formation.28 To seek a
saccharide part that could match triphenylalanine in terms of
interactions and spatial size and thus afford a parallel packing
model, we screened some typical oligosaccharides such as
lactose (Lac), trimannose (TMan), maltotriose (MalT), and
manner still remained stable after 50 ns relaxation but the
counterpart in an antiparallel manner tended to fragment
(Scheme 1c,d), whereas other glycopeptide models such as
Lac-3F, TMan-3F, and MalT-3F seemingly could not well
match triphenylalanine to afford a stable protofilament after
the same relaxation time (Schemes S1−S3). The outcome
suggested that the sialyllactose, a trisaccharide with a terminal
negative charge and an extended conformation (Scheme 1a)
appeared to be capable of matching triphenylalanine so as to
afford a desired glycopeptide model that preferably adopted
parallel packing as confirmed by the outperformance of parallel
packing to antiparallel packing in terms of saccharide H-bonds
and total H-bonds (Scheme 1e,f, Scheme S4).
SL-3F Associates into Twisted Double-Strand Fibrils.
SL-3F was synthesized via a coupling reaction between the
hydrazide group of triphenylalanine and the reducing end of
the sialyllactose (for details, see Schemes S5 and S6). Then the
compound was dispersed in water at 1 mg/mL. After several
6623
https://doi.org/10.1021/jacs.1c01951
J. Am. Chem. Soc. 2021, 143, 6622−6633

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
electron tomography (Figure S2). These double-strand fibrils
appeared to be uniform with respect to diameter (∼6.5 nm)
and pitch (∼198 nm) (Figure 1d,e). Very interestingly, these
double-strand fibrils have contour lengths up to hundreds of
microns (Figure 1f) yet seem to be capable of highly surviving
in solution after even 1 year of incubation as confirmed by the
unobservable slope variation in small-angle X-ray scattering
(SAXS) data (Figure S3a). The aspect ratio of these fibrils is as
high as 105 (Figure 1g), which seems larger than that of fibrous
populations built by synthetic small molecules even naturally
occurring biomacromolecules (Table S1). This extra-high
aspect ratio reveals the great stability of SL-3F building blocks
in the double strand fibrils, indicating our successful initiation
of designing SL-3F.
To understand the fibril association behavior, the formation
process of the double-strand fibrils was traced by means of
TEM and AFM (Figure 2). It revealed that micelles initially
formed (Figure 2a) then further fused into protofilaments
(Figure 2b), which would grow into single-strand protofibrils
(Figure 2c) with heights of 3.2 nm. Interestingly, these single-
strand protofibrils tended to laterally associate into ribbon-like
structure (as marked in Figure 2c), indicating their high
propensity with respect to lateral association. After 12 h
incubation, some single-strand protofibrils within ribbon-like
structures were found to associate into double-strand fibrils as
confirmed by the emergence of fibrous species with a height of
6.5 nm (Figure 2d, marked with red arrows). The remaining
single-strand protofibrils tended to complete the double-strand
process within 12 d once a small amount of double-strand
fibrils formed (Figure 2e−g). Cryo-EM, SAXS, circular
dichroism (CD), and UV−vis spectra were also used to
confirm the association processes of SL-3F mentioned above
(Figure S4−S6). The fibril association behaviors here reported,
namely, a single-strand protofibril initially formed and then
laterally associated into a ribbon-like structure where two
single-strand protofibrils intertwined into double-strand fibrils
Figure 1. Fibril association behavior of SL-3F (1 mg/mL of SL-3F in
water). (a, b, d, e) AFM height images and corresponding profiles of
double-strand fibrils, (c) negatively stained TEM image, (f) statistical
result of contour length of SL-3F, (g) a diagram showing aspect ratio
versus molecular weight of building blocks for various fibrous
assemblies (details in Table S1).
days, a fibrous structure was found in the solution. The
combination of transmission electron microscopy (TEM),
cryo-electron microscopy (cryo-EM), and atomic force
microscopy (AFM) proved that SL-3F gave rise to well-
defined fibrous structures (Figure 1a−c, Figure S1). These
resultant fibrous species were found to be left-handed and
double-stranded (Figure 1b,c), which was further evidenced by
Figure 2. Association processes of SL-3F as a function of incubation time; all scale bars are 50 nm (1 mg/mL of SL-3F in water).
6624
https://doi.org/10.1021/jacs.1c01951
J. Am. Chem. Soc. 2021, 143, 6622−6633

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 3. All-atom simulation for double-strand process. (a) relaxation process of SL-3F single-strand protofibril adopting parallel packing mode.
(b) Top view and front view of single-strand protofibril of SL-3F after blurring of the saccharide moiety. (c) Average statistical results of solvent-
accessible surface area (SASA) of each phenylalanine residue of SL-3F, (d) illustration for the competition between strand-forming and strand-
association. (e) Relaxation process of double-strand fibril of SL-3F. (f) SASA value of the phenylalanine residue contributed to strand association
before and after undergoing the double-strand process, (g) the statistical results of two types of H-bonds involved in double-strand fibrils of SL-3F,
(h) CD result of double-strand fibrils of SL-3F (1 mg/mL, water).
after a long induction period (Figure 2d), may provide the first
visible evidence of double-stranded process of fibrils.
residues tended to be embedded in the protofibril (marked as
yellow in Figure 3b), while a small amount of phenylalanine
residues (marked as green in Figure 3b) were still prone to
expose on the surface. The calculation of solvent accessible
surface area (SASA) allowed us to quantitatively identify the
role of each phenylalanine residue of a SL-3F molecule (Figure
3c). The SASA value of the middle phenylalanine residue (in
green in Figure 3d) was determined to be 0.90 nm2, but this
value for the two phenylalanine residues at both sides (in
yellow in Figure 3d) was only 0.43 and 0.40 nm2. This
quantified result clearly suggested that the two phenylalanine
residues at both sides that tended to be buried in the
protofibril would contribute to strand formation, while the
middle phenylalanine residue that was prone to expose on the
protofibril surface was most likely involved in strand
association (Figure 3d). Thereby four hydrophobic phenyl-
To rationalize this double-stranded process, molecular
dynamics simulation was performed to afford the detail at an
atomic level. As mentioned above, a starting structure in a
parallel manner (Figure 3a, left) was defined (details in Figure
S7), yet this amphiphilic structure may tend to form the initial
template of single-strand protofibril in a way that the two
hydrophobic ends of SL-3F close together (Figure 3a, middle)
to reduce the exposure of hydrophobic end to water. After 10
ns simulation, a stable single-strand protofibril with remarkable
left-handed feature emerged (Figure 3a, right).
Next, we attempted to rationalize the association of single-
strand protofibril into double-stranded fibril. After blurring the
saccharide moiety on the surface of a balanced single-strand
protofibril, it was readily found that most phenylalanine
6625
https://doi.org/10.1021/jacs.1c01951
J. Am. Chem. Soc. 2021, 143, 6622−6633

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
alanine residues appeared to engage in strand formation, but
only two hydrophobic phenylalanine residues would drive
single-strand protofibrils to laterally associate (Figure 3d).
That is, the driving force for strand formation seemed to be
larger than that for strand-association; in this case, strand
formation appeared to be preferred, accounting for the high
restraint of this type of protofibril with respect to lateral
association.
structure of SL-3F protofibrils may not be as regular as a β-
sheet. Expectedly, the combination of CD spectroscopy
(Figure 3h), Fourier transform infrared spectrum (FT-IR),
urea addition assay, and varying temperature CD assay (Figure
S10) proved that triphenylalanine within SL-3F protofibrils
indeed adopts a PP II helix conformation, which has a
relatively random amide arrangement in comparison to β-
sheet.32
Based the single-strand protofibril model, the formation of a
double-strand fibril was simulated starting from two laterally
located single-strand protofibrils, as indicated from our AFM
observation (Figure 2c,d). It was found that a stable double-
strand fibril structure with striking left-handed tendency was
yielded after a 10 ns relaxation on the initial template of
double-strand fibril (Figure 3e, supporting movie 1). The
simulation results showed that the average SASA of the middle
phenylalanine responsible for strand-association decreased
from 0.90 nm2 for a single-strand protofibril to 0.59 nm2 for
a double-strand fibril (Figure 3f), implying that the potential of
further lateral association of single-strand protofibril was
attenuated significantly by double-stranded association. In
addition, the all-atom simulation further suggested that only
double-strand fibrils came to emerge after certain relaxation
time even if more than two single-strand protofibrils
constituted the initial template (Figure S8). This result was
consistent with the aforementioned discussion, where the
driving force for strand-forming exceeded that for strand-
association and the single-strand protofibril preferentially
associated into double-strand species rather than underwent
infinite lateral association.
In short, the all-atom simulation explained the fibril
association very well since the obtained simulation structure
of the single-stranded protofibrils and double-stranded fibrils
were quite consistent with the experimental results. The
diameter of the single-strand protofibril of the simulation
model can be measured as 3.2 nm (Figure S9a), which is
consistent to the diameter measured by AFM as 3.2 nm
(Figure 2c). Similarly, the diameter of the double-stranded
fibril in simulation was measured as 6.4 nm (Figure S9b),
which is consistent with the diameter of 6.5 nm as observed by
AFM (Figure 1d). More importantly, starting from two
laterally associated single-strand protofibrils as observed in
experiments (Figure 2d), the left-handed twisting tendency of
the double-stranded fibril was very similar to the left-
handedness found under AFM (Figure 1b) and TEM (Figure
1c). Thus, we may conclude that the all-atom simulation
explained the observed fibril association very well, providing us
a whole picture on the molecular level from the starting
structure of SL-3F to single-strand protofibril, then to left-
handed double-strand fibrils. The consistency of simulation
and experiments gave us an opportunity to connect the
molecular structure to the various fibrous behaviors.
A relatively random amide arrangement arising from PP II
helix conformation was thought to be capable of removing
certain restrictions on the spatial motion of hydrophobic
phenylalanine residues,33 allowing the moderate exposure of
some hydrophobic residues on the surface of protofibrils as
confirmed by the determined SASA value of 0.90 nm2 of each
middle phenylalanine residue. Meanwhile, the carboxyl groups
of sialyllactose exposed on the surface of protofibrils tended to
be partially deprotonated in a neutral environment, giving rise
to electrostatic repulsion interactions between single-strand
protofibrils. Thus, hydrophobic and electrostatic repulsion
interactions appeared to be capable of antagonistically defining
the association behavior of single-strand protofibrils. It was
found that the double-stranding process tended to cause an
increase in the negative charge on the fibril surface as
confirmed by zeta-potential measurement (Figure S11a) and
also give rise to a 30% decrease in the exposed area of
hydrophobic residue on the fibril surface (Figure 3f). As the
electrostatic repulsion force emerged as a dominant interaction
for fibril association, these double-strand fibrils tended to
remain highly stable rather than continue to laterally associate.
We also experimentally investigated the interfibril hydrophobic
interactions. The outcome showed that the process of two
single-stranded protofibril associating into a double-stranded
fibril is often accompanied by appreciable changes in
hydrophobic interactions (Figure S11b), suggesting the
involvement of hydrophobic interactions in the fibril
association process.
As an intermediate species of particular interest, the ribbon-
like structure involved in the association process of SL-3F
(Figure S12a) could be rationalized on the basis of the
compromise of electrostatic repulsion to hydrophobic
interaction among SL-3F single-strand protofibrils. To further
verify this point, 1 μL of NaOH solution (0.5 M) was added
into the 1 mL solution of the ribbon-like structure to
strengthen the electrostatic repulsion force among single-
stranded protofibrils. After only 0.5 h, the ribbon-like structure
already became relatively discrete, and then these discrete
single-strand protofibrils further associated into ultralong
double-strand fibrils within 8 h (Figure S12). The speed of
fibril association in basic conditions appeared to be dozens of
times faster than that in a neutral environment (Figure S13),
verifying the strong contribution from electrostatic repulsion.
Taken together, it appeared that this type of glycopeptide
model synergistically encoded the fibril association behavior
through two key parameters. One parameter referred to the
competition between the strand-forming tendency and strand-
association tendency, which was encoded by the peptide
moiety; another parameter is the competitive set of
interactions including electrostatic repulsion and hydrophobic
interaction that resided together on the surface of protofibrils,
which appeared to be encoded by the saccharide moiety
because the saccharide moiety not only contributed to the
electrostatic repulsion but also indirectly mediated the
exposure of hydrophobic residues on the surface of protofibrils
On the basis of the aforementioned results, the exact role of
the saccharide part was further identified by simulation. It
appeared that two sets of hydrogen bonds were involved in the
protofibrils of SL-3F (Scheme S4), including saccharide H-
bonds that were prone to be unstructured as a result of their
steric nature29 and peptide H-bonds that having high
propensity for forming β-sheet.30,31 The quantitative compar-
ison revealed that saccharide H-bonds were slightly more than
peptide H-bonds (Figure 3g), allowing the unstructured
saccharide H-bonds, to some extent, to perturb the arrange-
ment of regular peptide H-bonds. Thus, the secondary
6626
https://doi.org/10.1021/jacs.1c01951
J. Am. Chem. Soc. 2021, 143, 6622−6633

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 4. Fibril association behavior of SL-4F (1 mg/mL of SL-4F in water). (a, b) Negatively stained TEM (c) cryo-EM images of twisted ribbon
structure of SL-4F, (d) top view and front view of a single-strand protofibril of SL-4F after blurring the saccharide moiety, (e) statistical results of
the solvent-accessible surface area (SASA) of each phenylalanine residue of SL-4F, (f) illustration for the competition between strand-forming and
strand-association, (g) the AFM height image of radially associated protofibrils of SL-4F, (h) relaxation process of twisted ribbon of SL-4F (front
view and side view), (i) the statistical results of two types of H-bonds involved in protofibrils of SL-4F, (j) the CD result of SL-4F (1 mg/mL,
water); scale bar in a, c, and g is 50, 50, and 10 nm, respectively.
by affecting the secondary structure of protofibrils due to their
involvement in the H-bond competition with the peptide
moiety.
twisted nanoribbon structure with width centered at 25 nm
and pitch of 280 nm (Figure 4a-c, S15). Excitingly, these well-
defined nanoribbons were found to be right-handed (Figure
4b), indicating that SL-4F gave rise to a new fibril association
behavior. Molecular dynamics simulation was performed again
to rationalize this change. Starting from the parallel packing of
SL-4F as SL-3F, the model of single-strand protofibril was
built again by twisting two of the starting structures together
(Figure S16a, simulation details in SI). The quantified results
showed that, in the case of four hydrophobic phenylalanine
residues of SL-4F, two (marked as yellow in Figure 4e, 4f) with
SASA of 0.15and 0.30 nm2 would contribute to strand-forming
and another two (marked as green in Figure 4e,f) with SASA of
1.31 and 0.51 nm2 would engage in strand-association (Figure
4d−f). Thus, the driving force for strand-forming (four
phenylalanine residues) could be equal to that for strand-
association (four phenylalanine residues) (Figure 4f); in this
case, the strand-association might severely interfere the
occurrence of strand-forming event.
Moreover, the involvement of two phenylalanine residues
with SASA values up to 1.31 and 0.51 nm2 into strand-
association would cause a strong hydrophobic interaction
among single-strand protofibrils of SL-4F, which would
facilitate the radial association of the protofibrils while severely
inhibiting their axial growth, giving rise to the radial association
of short protofibrils into a ribbon structure (Figure S16b).
Indeed, the AFM image showed that the SL-4F afforded only
short protofibrils with high propensity to undergo radial
association (Figure 4g, Figure S17). According to this
experimental result, we defined the initial template of the
To further unravel the contribution of the saccharide moiety
on the fibril association behaviors, we selected to remove the
sialic acid from sialyllactose in situ by sialidase to obtain Lac-
3F. As shown in Figure S14c, the removal of sialic acid
triggered the switch from ultralong fibrils into vesicles. The
structural evolution process was traced by TEM; initially, the
long fibrils collapsed into short fibrils, and then these short
fibrils tended to laterally associate into bundles and further
fused into nanosheets, which would spontaneously curl into
vesicles in solution (Figure S14e). The decrease of aggregate
size with incubation time observed in DLS assay appeared to
well match this structural evolution (Figure S14f). More
importantly, the conversion of secondary structure from PP II
helix to β-sheet (Figure S14d) suggested that the removal of
sialic acid significantly weakened the interference of unstruc-
tured saccharide H-bonds on regular peptide H-bonds; as a
result, the intrinsic priority of forming a β-sheet of a peptide
moiety tended to be liberated. This result appeared to
exemplify the effect of the subtle trade-off of the saccharide
moiety and the peptide part on the fibril association behavior.
SL-4F Radially Associates into a Twistedly Right-
Handed Nanoribbon. To further determine the synergistic
definition of peptides and saccharides on the fibril association
behaviors, SL-4F (details in Scheme S6) was employed as
another model to highlight the contribution from peptides via
addition of one phenylalanine. By applying the same assembly
scenario of SL-3F, it was found that SL-4F associated into a
6627
https://doi.org/10.1021/jacs.1c01951
J. Am. Chem. Soc. 2021, 143, 6622−6633

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 5. Fibril association behavior of SL-2F (1 mg/mL of SL-2F in water). (a, b) AFM height images, (c) Cryo-EM image of multistrand ribbon
structure of SL-2F, (d) all-atom simulation result of a single-strand protofibril, (e) top view and front view of single-strand protofibrils of SL-2F
after blurring the saccharide moiety, (f) the statistical results of solvent accessible surface area (SASA) of each phenylalanine residues of SL-2F, (g)
rheological behavior of SL-2F gel (0.1 mass%), (h) statistical results of two types of H-bonds involved in protofibrils of SL-2F, (i) CD result of SL-
2F (1 mg/mL, water), (j) illustration for the competition between strand-forming and strand-association; scale bar in a, b, and c is 100 nm.
nanoribbon (Figure 4h). After 10 ns simulation, along the left-
handedness tendency that emerged in each single-strand
protofibril (as marked with dotted box in Figure 4h), the
entire ribbon structure yet displayed a pronounced right-
handed tendency (labeled with red dot in Figure 4h,
supporting movie 2). It was supposed that the molecular
chirality of phenylalanine triggered the axial left-handedness of
each single-strand protofibril,34 and the twisting to the left of
the single-strand protofibrils that adhered to each other via
strong hydrophobic interaction may in turn drive the entire
ribbon to twist to the right.35
Notably, in the case of SL-4F, the quantitative excess of
peptide H-bonds to saccharide H-bonds (Figure 4i) possibly
liberated the high propensity for β-sheet formation of the
peptide moiety as confirmed by the CD results (Figure 4j). In
general, the all-atom simulation illustrated the fibril association
of SL-4F very well since the obtained simulation structure of
the single-stranded protofibrils and twisted nanoribbons
seemed to be quite consistent with the experimental results.
The diameter of the single-stranded protofibril of the
simulation model can be measured as 3.7 nm (Figure S17d),
which was consistent with the diameter measured by AFM as
3.7 nm (Figure S17c). More importantly, starting from some
radially associated single-stranded protofibrils as observed in
experiments (Figure 4g), the right-handed twisting tendency of
the nanoribbon was very similar to the right-handedness found
under TEM (Figure 4b,c).
SL-2F Laterally Associates into Multistrand Ribbon
and Hydrogel. To gain further insights to the encoding of
saccharides and peptides on the fibril association, SL-2F was
exploited as another model to emphasize the contribution from
saccharide via removal of one phenylalanine. The assembly of
SL-2F was prepared following the same procedure as for SL-
3F and SL-4F. The combination of AFM and Cryo-EM results
showed that SL-2F laterally associated into a multistrand
ribbon comprising single-strand protofibrils as confirmed by a
6628
https://doi.org/10.1021/jacs.1c01951
J. Am. Chem. Soc. 2021, 143, 6622−6633

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 6. (a) Statistical results of two sets of H-bonds involved in protofibrils and (b) rheological behavior (0.5 mass%) of diverse glycopeptides.
The saccharide H-bond/peptide H-bond (S/P) calculated from the averaged saccharide H-bonds within the simulation time was divided by the
averaged peptide H-bonds within the simulation time. (c) Hydrogel strength (G′) as a function of number ratios of saccharide H-bonds to peptide
H-bonds (S/P).
height of about 3.5 nm (Figure 5a−c). The trace on the
association process showed that in the initial 12 h SL-2F
appeared to share a similar association process with SL-3F,
such as micelles, protofilaments, and multistrand ribbon
sequentially formed (Figure S18). Unexpectedly, after a 48 h
incubation, these single-strand protofibrils tended to undergo
an infinitely lateral association to afford a hydrogel with a
superhigh storage modulus up to 19000 Pa, rather than
associate into highly discrete double-strand fibrils like SL-3F
(Figure 5a,g, Figure S19).
To rationalize what differentiated the fibril association
behaviors of SL-2F and SL-3F, molecular dynamics simulation
was performed again. The aforementioned parallel packing of
SL-2F, like those of SL-3F/SL-4F, was employed, which was
further associated into a single-strand protofibril. Then a stable
single-strand protofibril structure with a pronounced left-
handed tendency was yielded after 50 ns simulation (Figure
5d, Figure S20). Notably, the diameter of a stable single-strand
protofibril was determined to be about 3.5 nm (Figure S20a),
which was consistent to the observation from an AFM image
(Figure 5b). After the saccharide moiety was blurred on the
surface of a single-strand protofibril, it appeared that the
exposure of hydrophobic phenylalanine residues on the surface
of SL-2F single-strand protofibrils was more significant than
that on the surface of SL-3F (Figure 5e, Figure 3b). The
quantified results showed that the average SASA value of each
phenylalanine residue exposed on the protofibril surface of SL-
2F was 1.25 nm2 (Figure 5f), which was nearly 40% higher
than that of SL-3F (0.90 nm2), implying a strong hydrophobic
interaction among protofibrils of SL-2F. This enhanced
hydrophobic interaction could be illustrated from the
remodeled synergy of the saccharide part and peptide moiety.
In detail, in the case of SL-2F, the saccharide H-bonds were
twice as high as peptide H-bonds (Figure 5h), allowing the
unstructured saccharide H-bonds to seriously interfere with the
arrangement of regular peptide H-bonds.36 Thus, the
secondary structure adopted by diphenylalanine within SL-
2F protofibrils may be irregular. Indeed, a prominent
maximum at 220 nm arising from phenylalanine chromo-
phores37 indicated that diphenylalanine within SL-2F proto-
fibrils adopted a random packing mode (Figure 5i). It was
thought that random amide arrangement hardly imposed
adequate restrictions on the spatial motion of hydrophobic
phenylalanine residues, thus giving rise to a large exposure of
hydrophobic residues on the surface of protofibrils.33
Notably, for the two phenylalanine residues of SL-2F, one
with SASA of 0.46 nm2 would contribute to strand-forming
and another one with SASA of 1.25 nm2 would engage in
strand-association; thus, the driving force for strand-forming
(two phenylalanine residues) was equal to that for strand-
association (two phenylalanine residues) (Figure 5j). Similar
to SL-4F, the strand-association might severely interfere
strand-forming event, thus hardly giving rise to long fibrous
structures. On the basis of the above discussion, we intend to
believe that the dominant strand-association and the relatively
strong hydrophobic interaction drove the single-strand
protofibrils of SL-2F to undergo infinitely lateral association
and finally to gel.
We also replaced the middle phenylalanine for strand-
association of SL-3F with either tyrosine to obtain SL-FYF,
6629
https://doi.org/10.1021/jacs.1c01951
J. Am. Chem. Soc. 2021, 143, 6622−6633

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 7. (a) Schematic illustration for transwell assay, (b) in vitro inhibition effect of various fibrils on B16F10 cells migration, (c) in vitro
inhibition of SL-3F double-strand fibrils on the growth of B16F10 cells, visualization of interaction between SL-3F double-strand fibrils and
B16F10 cells by (d) CLSM and (e) FE-SEM. Data are presented as mean s.e.m. Statistical significance was calculated by Student’s t test: ***P <
0.001, **P < 0.01, *P < 0.05.
which may contribute to lower hydrophobic interaction than
phenylalanine, or fluorophenylalanine to obtain SL-FUF,
which could provide higher hydrophobic interaction than
phenylalanine (Scheme S6). SL-FYF was found to associate
into discrete single-strand protofibrils rather than double-
strand fibers like SL-3F, possibly because of the weak driving
force for strand-association (Figure S21a). In the case of SL-
FUF, the most frequently occurring population appeared to be
a thick fibrous structure composed of double-strand fibrils with
6.3 nm height (Figure S21j), displaying a higher tendency of
lateral association than that of SL-FYF and SL-3F, which could
be ascribed to the stronger driving force for strand-association.
Correlation between Molecular Building Blocks and
Hydrogelation Properties. Having dissected the relation-
ship between molecular building blocks and fibril association
behaviors, next we tentatively leveraged the fibril association
behaviors to bridge the molecular building blocks with some
macroscopic properties in terms of hydrogelation. As
mentioned above, the trade-off of saccharide H-bonds and
peptide H-bonds demonstrated a critical determinant for fibril
association; this inspired us to further correlate these two sets
of H-bonds with hydrogelation properties of glycopeptide
fibrils. We first dispersed five types of glycopeptides into water
under neutral conditions with an identical concentration of 0.5
mass % and then incubated for 12 d. On the basis of the
quantitative comparison of two sets of H-bonds as well as the
hydrogel strength of five types of glycopeptides, an appreciable
relationship between these two aspects emerged (Figure 6a,b).
To better understand this relationship, we further introduced a
parameter S/P, i.e., number ratio of saccharide H-bonds to
peptide H-bonds to build a bridge between molecular building
blocks and hydrogel strength (Figure 6c).
renders these fibrils self-limiting with respect to infinitely
lateral association and bundle formation that often gives rise to
hydrogel.38,39 When the S/P ratio was increased to a higher
value (1.4 < S/P < 1.9), more unstructured saccharide H-
bonds might have interfered with the regular packing of
peptide part and thus enabled larger exposure of the
fluorophenylalanine residue of SL-FUF compared to SL-3F.
Thus, the double-strand fibrils with a strong hydrophobic
surface (as marked with light orange in Figure 6c) tended to
associate into thick fibrils providing a weak hydrogel with
modulus of 520 Pa. Interestingly, for SL-FYF, the flexible
single-strand protofibrils defined by dominant unstructured
saccharide H-bonds and weak driving force for strand-
association afforded a hydrogel with a modulus of 2100 Pa,
and the most likely reason could be the strong topological
interactions among flexible single-strand protofibrils that
hardly underwent any further association.40 As the S/P ratio
was more than 1.9, the fully dominant saccharide H-bonds
imparted the single-strand protofibrils of SL-2F with very
strong hydrophobic surface (as marked with orange in Figure
6c), these protofibrils tended to undergo finitely lateral
association to afford a very strong hydrogel with modulus up
to 32500 Pa.
The saccharide moiety, on one hand, engaged in the
modulation of the hydrogelation properties with the form of
saccharide H-bonds; on the other hand, the electrostatic
interaction provided by the saccharide moiety could also
manipulate the hydrogelation properties. For example, the
addition of a small amount of HCl (1 μL, 0.5 M) would trigger
the highly stable fibrils of SL-3F to collapse into hydrogel with
a modulus of 400 Pa (Figure S22), and possibly the weakened
surface negative charge gave rise to the collapse of SL-3F
fibrils. In short, we suppose that the hydrogelation property
was still governed by the saccharide moiety and peptide part.
Correlation between Fibril Association Behaviors
and Antitumor Activities. Given that the fibrous popula-
tions investigated here exhibited a high structural similarity to
extracellular matrices, a niche can regulate tumor invasion and
migration via cell−matrix interaction.41 Moreover, the natural
tumor microenvironments often impart changes in cell−matrix
interaction through variation with respect to fibril association
Fortunately, a clear relationship between the molecular
building blocks and hydrogel strength can be displayed as
follows. At low S/P ratios (<1.4), the dominant peptide H-
bonds tended to enable the glycopeptides assembling into
twisted fibrils, such as twisted double-strand fibrils of SL-3F
and twisted nanoribbons of SL-4F. Notably, these twisted
fibrils were found to be capable of good survival in aqueous
solution and therefore always afforded solutions with no
appreciable modulus, possibly since the twisted structure
6630
https://doi.org/10.1021/jacs.1c01951
J. Am. Chem. Soc. 2021, 143, 6622−6633

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
behavior.42 All of these inspired us to correlate fibril
association behavior with cell−matrix interactions. Thus, we
selected diverse tumor cells for rationalizing this correlation on
the basis of the presence of cell−matrix interactions between
overexpressed sialyllactose receptors on tumor cell surfaces and
sialyllactose exposed on the fibrils surface.43
of glycopeptide models. The subtle trade-off of oligosaccharide
and oligopeptide components was proven to be a determinant
that regulates the growth direction as well as chirality of fibrils.
The fibril association was also identified as a useful tool to
bridge the molecular building blocks and some macroscopic
properties in terms of hydrogelation as well as antitumor
activities. Notably, the double-strand fibrils of SL-3F with a
super high aspect ratio up to 105 were found to be capable of
highly surviving in solution and firmly sticking on the tumor
cells surface. Glycosylation is a ubiquitous phenomenon in
Nature, which is involved in various fibrous association
processes with distinct effects. We hope that the molecular
details of saccharides on fibrous association may reveal some
mechanisms involved by glycans and predict unraveled
disciplines. In short, the reported comprehensive landscape
of fibril association behaviors may provide tools to decode
naturally occurring fibril association and shed unique insight
on fibrous material design.
Tumor migration assay performed by the transwell method
was selected to evaluate the effect of fibril association behavior
on cell−matrix interaction (Figure 7a). These five types of
fibrous populations exhibited no detected cytotoxicity even
when the concentration was up to 0.25 mg/mL (Figure S23).
The diverse nongelling fibrous species were incubated with
B16F10 cells at a concentration of 0.125 mg/mL. The in vitro
outcomes showed that the double-strand fibrils of SL-3F and
single-strand fibrils of SL-FYF could display certain inhibition
on the migration of B16F10 cells while double-strand fibrils of
SL-3F exhibited the most effective inhibition, but the fibrous
species of SL-2F, SL-4F, or SL-FUF could not (Figure 7b,
Figure S24). The universality of the double-strand fibrils of SL-
3F to inhibit migration of diverse tumor cells such as BV-2 and
HepG 2 is also demonstrated in the Supporting Information
(Figure S25). We also tentatively quantified interactions
between SL-3F fibrils and B16F10 cells, and the outcome
showed that each microgram fibrils of SL-3F could adhere (2.8
0.5) × 103 B16F10 cells (Figure S26).
The double-strand fibrils of SL-3F were also proven to be
capable of delaying the growth of tumor cells (Figure 7c). To
further rationalize the mechanism underlying inhibition of
tumor cell migration and growth, the visualization of the
interaction between distinct fibrils and tumor cells was
performed. Co-assembly of SL-3F or SL-FYF with FITC-3F
afforded the corresponding fluorescence-labeled fibrils, which
were named as SL-3F+FITC-3F and SL-FYF+FITC-3F,
respectively. When these fluorescence-labeled fibrils were
cocultured with B16F10 cells for 48 h, SL-3F+FITC-3F
appeared to tightly stick to the surface of B16F10 cells like
molecular glue44 (Figure 7d), whereas neither SL-FYF+FITC-
3F nor only FITC-3F could effectively stick to the surface of
B16F10 cells (Figure S27), suggesting that the capability of
sticking cells together of the SL-3F fibrils might be one of the
factors that can inhibit tumor migration. The FE-SEM images
provided high resolution evidence of interaction between
double-strand fibrils and tumor cells (Figure 7e).
ASSOCIATED CONTENT
■
sı
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/jacs.1c01951.
Synthesis, prescreening of glycopeptide modes, charac-
terization of fibrils, and antitumor activities of the fibrils
(PDF)
Stable double-strand fibril structure with striking left-
handed tendency (MPG)
Nanoribbon with pronounced right-handed tendency
(MPG)
AUTHOR INFORMATION
■
Corresponding Author
Guosong Chen − The State Key Laboratory of Molecular
Engineering of Polymers and Department of Macromolecular
Science and Multiscale Research Institute of Complex
Systems, Fudan University, Shanghai 200433, P.R. China;
orcid.org/0000-0001-7089-911X; Email: guosong@
fudan.edu.cn
Authors
The correlation between fibril association behaviors and
antitumor activities was further interrogated by in vivo assay.
SL-FYF and SL-3F, two fibrous species with similar
morphology of high aspect ratio but distinct association
behavior (the former is single-stranded, the latter is double-
stranded), was selected to perform the in vivo assay. We first
injected B16F10 cells into the abdominal cavity of the mice
and then injected the fibrous species of SL-FYF and SL-3F at
the same location of the tumor. After 6 d incubation, the
antitumor activities of SL-3F seemed to outperform that of SL-
FYF (Figure S28), which was well consistent with the in vitro
results (Figure 7b). This difference in antitumor activity
seemly emerged from the distinct capabilities of sticking tumor
cells of fibrous species, which was significantly influenced by
their fibril association behaviors.
Rongying Liu − The State Key Laboratory of Molecular
Engineering of Polymers and Department of Macromolecular
Science, Fudan University, Shanghai 200433, P.R. China
Ran Zhang − State Key Laboratory of Polymer Physics and
Chemistry, Changchun Institute of Applied Chemistry,
Chinese Academy of Sciences, Changchun 130022, P.R.
China
Long Li − The State Key Laboratory of Molecular Engineering
of Polymers and Department of Macromolecular Science,
Fudan University, Shanghai 200433, P.R. China
Zdravko Kochovski − Department for Electrochemical Energy
Storage, Helmholtz-Zentrum Berlin fu
Energie, 14109 Berlin, Germany
̈
r Materialien und
Lintong Yao − The State Key Laboratory of Molecular
Engineering of Polymers and Department of Macromolecular
Science, Fudan University, Shanghai 200433, P.R. China
Mu-Ping Nieh − Polymer Program, Institute of Materials
Science and Department of Chemical and Biomolecular
Engineering, University of Connecticut, Storrs, Connecticut
06269, United States; orcid.org/0000-0003-4462-8716
CONCLUSION
■
In summary, we have experimentally and theoretically
demonstrated the synergistic encoding of saccharides and
peptides on fibril association behaviors based on a designed set
6631
https://doi.org/10.1021/jacs.1c01951
J. Am. Chem. Soc. 2021, 143, 6622−6633

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(11) Freeman, R.; et al. Reversible self-assembly of superstructured
networks. Science 2018, 362, 808−813.
(12) Jones, C. D.; Simmons, H. T. D.; Horner, K. E.; Liu, K.;
Thompson, R. L.; Steed, J. W. Braiding, branching and chiral
amplification of nanofibres in supramolecular gels. Nat. Chem. 2019,
11, 375−381.
(13) Khurana, R.; Ionescu-Zanetti, C.; Pope, M.; Li, J.; Nielson, L.;
Ramirez-Alvarado, M.; Regan, L.; Fink, A. L.; Carter, S. A. A general
model for amyloid fbril assembly based on morphological studies
using atomic force microscopy. Biophys. J. 2003, 85, 1135−1144.
(14) Ridgley, D. M.; Barone, J. R. Evolution of the amyloid fiber
over multiple length scales. ACS Nano 2013, 7, 1006−1015.
(15) Hennet, T. Collagen glycosylation. Curr. Opin. Struct. Biol.
2019, 56, 131−13.
(16) Lewis, Y. E.; Galesic, A.; Levine, P. M.; De Leon, C. A.; Lamiri,
N.; Brennan, C. K.; Pratt, M. R. O-GlcNAcylation of α-synuclein at
serine 87 reduces aggregation without affecting membrane binding.
ACS Chem. Biol. 2017, 12, 1020−1027.
(17) Signas, C.; et al. Nucleotide sequence of the gene for a
fibronectin-bindng protein from Staphylococcus aureus: use of this
peptide seuence in the synthesis of biologically active peptides. Proc.
Natl. Acad. Sci. U. S. A. 1989, 86, 699−703.
(18) Marotta, N. P.; Lin, Y. H.; Lewis, Y. E.; Ambroso, M. R.; Zaro,
B. W.; Roth, M. T.; Pratt, M. R. O-GlcNAc modification blocks the
aggregation and toxicity of the protein α-synuclein associated with
Parkinson’s disease. Nat. Chem. 2015, 7, 913−920.
(19) Levine, P. M.; Galesic, A.; Balana, A. T.; Mahul-Mellier, A. L.;
Navarro, M. X.; De Leon, C. A.; Pratt, M. R. α-Synuclein O-
GlcNAcylation alters aggregation and toxicity, revealing certain
residues as potential inhibitors of Parkinson’s disease. Proc. Natl.
Acad. Sci. U. S. A. 2019, 116, 1511−1519.
(20) Galesic, A.; Rakshit, A.; Cutolo, G.; Pacheco, R. P.; Balana, A.
T.; Moon, S. P.; Pratt, M. R. Comparison of N-Acetyl-Glucosamine to
Other Monosaccharides Reveals Structural Differences for the
Inhibition of α-Synuclein Aggregation. ACS Chem. Biol. 2021, 16,
14−19.
(21) Reches, M.; Gazit, E. Casting metal nanowires within discrete
self-assembled peptide nanotubes. Science 2003, 300, 625−627.
(22) Brown, N.; Lei, J.; Zhan, C.; Shimon, L. J. W.; Adler-
Abramovich, L.; Wei, G.; Gazit, E. Structural polymorphism in a self-
assembled tri-aromatic peptide system. ACS Nano 2018, 12, 3253−
3262.
(23) Fleming, S.; Ulijn, R. V. Design of nanostructures based on
aromatic peptide amphiphiles. Chem. Soc. Rev. 2014, 43, 8150−8177.
(24) Pappas, C. G.; et al. Dynamic peptide libraries for the discovery
of supramolecular nanomaterials. Nat. Nanotechnol. 2016, 11, 960−
967.
(25) Iconomidou, V. A.; Hamodrakas, S. J. Natural protective
amyloids. Curr. Protein Pept. Sci. 2008, 9, 291−309.
(26) Gelenter, M. D.; et al. The peptide hormone glucagon forms
amyloid fibrils with two coexisting β-strand conformations. Nat.
Struct. Mol. Biol. 2019, 26, 592−598.
(27) Maji, S. K.; Drew, M. G.; Banerjee, A. First crystallographic
signature of amyloid-like fibril forming β-sheet assemblage from a
tripeptide with non-coded amino acids. Chem. Commun. 2001, 1946−
1947.
(28) Kemper, B.; Zengerling, L.; Spitzer, D.; Otter, R.; Bauer, T.;
Besenius, P. Kinetically controlled stepwise self-assembly of AuI-
metallopeptides in water. J. Am. Chem. Soc. 2018, 140, 534−537.
(29) van de Bovenkamp, F. S.; Hafkenscheid, L.; Rispens, T.;
Rombouts, Y. The emerging importance of IgG Fab glycosylation in
immunity. J. Immunol. 2016, 196, 1435−1441.
(30) Ma, X.; Wang, Y. Q.; Hua, J. A.; Xu, C. Y.; Yang, T.; Yuan, J.;
Chen, G. Q.; Guo, Z. J.; Wang, X. Y. A β-sheet-targeted theranostic
agent for diagnosing and preventing aggregation of pathogenic
peptides in Alzheimer’s disease. Sci. China: Chem. 2020, 63, 73−82.
(31) Yin, Q.; Liu, J. J.; Tian, M. T.; Xie, H.; Shen, L.; Sun, T. L.
Protein Fibrillation in Neurodegenerative Diseases and Its Chiral
Interaction with Interfaces. Acta Polym. Sin. 2019, 50, 575−587.
Yan Lu − Department for Electrochemical Energy Storage,
Helmholtz-Zentrum Berlin fur Materialien und Energie,
̈
14109 Berlin, Germany; Institute of Chemistry, University of
Potsdam, 14476 Potsdam, Germany; orcid.org/0000-
0003-3055-0073
Tongfei Shi − State Key Laboratory of Polymer Physics and
Chemistry, Changchun Institute of Applied Chemistry,
Chinese Academy of Sciences, Changchun 130022, P.R.
China; orcid.org/0000-0002-6763-2200
Complete contact information is available at:
https://pubs.acs.org/10.1021/jacs.1c01951
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
G.C. thanks NSFC/China (Nos. 51721002, 21861132012,
91956127, 21975047) for financial support. R.Z. thanks
NSFC/China (No. 21674114, 91956127) for financial
support. Y.L. thanks the Deutsche Forschungsgemeinschaft
(DFG, German Research Foundation)Project No.
410871749 for financial support. The authors thank the Joint
Lab for Structural Research at the Integrative Research
Institute for the Sciences (IRIS Adlershof, Berlin) for Cryo-
EM imaging. This work is supported by the Shanghai
Municipal Science and Technology Major Project
(No.2018SHZDZX01) and ZJ Lab.
REFERENCES
■
(1) Lieleg, O.; Claessens, M. M.; Bausch, A. R. Structure and
dynamics of cross-linked actin networks. Soft Matter 2010, 6, 218−
225.
(2) Marino, M.; Vairo, G. Multiscale Elastic Models of Collagen Bio-
structures: From Cross-Linked Molecules to Soft Tissues. Multiscale
Computer Modeling in Biomechanics and Biomedical Engineering 2013,
14, 73.
(3) Lim, C. Z. J.; et al. Subtyping of circulating exosome-bound
amyloid β reflects brain plaque deposition. Nat. Commun. 2019, 10, 1.
(4) Baker, M. B.; Albertazzi, L.; Voets, I. K.; Leenders, C. M. A.;
Palmans, A. R. A.; Pavan, G. M.; Meijer, E. W. Consequences of
chirality on the dynamics of a water-soluble supramolecular polymer.
Nat. Commun. 2015, 6, 6234.
(5) Yano, K.; Hanebuchi, T.; Zhang, X. J.; Itoh, Y.; Uchida, Y.; Sato,
T.; Matsuura, K.; Kagawa, F.; Araoka, F.; Aida, T. Supramolecular
Polymerization in Liquid Crystalline Media: Toward Modular
Synthesis of Multifunctional Core - Shell Columnar Liquid Crystals.
J. Am. Chem. Soc. 2019, 141, 10033−10038.
(6) Weyandt, E.; ter Huurne, G. M.; Vantomme, G.; Markvoort, A.
J.; Palmans, A. R. A.; Meijer, E. W. Photodynamic Control of the
Chain Length in Supramolecular Polymers: Switching an Intercalator
into a Chain Capper. J. Am. Chem. Soc. 2020, 142, 6295−6303.
(7) Bera, S.; Mondal, S.; Xue, B.; Shimon, L. J. W.; Cao, Y.; Gazit, E.
Rigid helical-like assemblies from a self-aggregating tripeptide. Nat.
Mater. 2019, 18, 503−509.
(8) Bera, S.; Xue, B.; Rehak, P.; Jacoby, G.; Ji, W.; Shimon, L. J. W.;
Beck, R.; Kral, P.; Cao, Y.; Gazit, E. Self-Assembly of Aromatic Amino
Acid Enantiomers into Supramolecular Materials of High Rigidity.
ACS Nano 2020, 14, 1694−1706.
(9) Hendrikse, S. I. S.; Su, L.; Hogervorst, T. P.; Lafleur, R. P. M.;
Lou, X. W.; Marel, G. A. V. D.; Codee, J. D. C.; Meijer, E. W.
Elucidating the Ordering in Self-Assembled Glycocalyx Mimicking
Supramolecular Copolymers in Water. J. Am. Chem. Soc. 2019, 141,
13877−13886.
(10) Adamcik, J.; et al. Understanding amyloid aggregation by
statistical analysis of atomic force microscopy images. Nat. Nano-
technol. 2010, 5, 423−428.
6632
https://doi.org/10.1021/jacs.1c01951
J. Am. Chem. Soc. 2021, 143, 6622−6633

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(32) Bochicchio, B.; Tamburro, A. M. Polyproline II structure in
proteins: identification by chiroptical spectroscopies, stability, and
functions. Chirality 2002, 14, 782−792.
(33) Ortony, J. H.; Newcomb, C. J.; Matson, J. B.; Palmer, L. C.;
Doan, P. E.; Hoffman, B. M.; Stupp, S. I. Internal dynamics of a
supramolecular nanofibre. Nat. Mater. 2014, 13, 812−816.
(34) Yue, B. B.; Zhu, L. L. Dynamic modulation of supramolecular
chirality driven by factors from internal to external level. Chem. - Asian
J. 2019, 14, 2172−2180.
(35) Efrati, E.; Irvine, W. T. M. Orientation-dependent handedness
and chiral design. Phys. Rev. X 2014, 4, 011003.
(36) Schoenmakers, S. M. C.; Leenders, C. M. A.; Lafleur, R. P. M.;
Lou, X. W.; Meijer, E. W.; Pavan, G. M.; Palmans, A. R. A. Impact of
the water-compatible periphery on the dynamic and structural
properties of benzene-1,3,5-tricarboxamide based amphiphiles.
Chem. Commun. 2018, 54, 11128−11131.
(37) Hamley, I. W.; Krysmann, M. J.; Kelarakis, A.; Castelletto, V.;
Noirez, L.; Hule, R. A.; Pochan, D. J. Nematic and columnar ordering
of a PEG-peptide conjugate in aqueous solution. Chem. - Eur. J. 2008,
14, 11369−11375.
(38) Marty, R.; Nigon, R.; Leite, D.; Frauenrath, H. Two-Fold odd -
even effect in self-assembled nanowires from oligopeptide-polymer-
substituted perylene bisimides. J. Am. Chem. Soc. 2014, 136, 3919−
3927.
(39) Cui, C. Y.; Chen, X. Y.; Liu, B.; Wu, T. L.; Fan, C. C.; Liu, W.
G. A High Strength Instant Adhesive Nano-hybrid Hydrogel as First-
aid Bandage. Acta Polym. Sin. 2019, 50, 613−622.
(40) Raghavan, S. R.; Douglas, J. F. The conundrum of gel formation
by molecular nanofibers, wormlike micelles, and filamentous proteins:
gelation without cross-links? Soft Matter 2012, 8, 8539−8546.
(41) Theocharis, A. D.; Skandalis, S. S.; Gialeli, C.; Karamanos, N.
K. Extracellular matrix structure. Adv. Drug Delivery Rev. 2016, 97, 4−
27.
(42) Walters, N. J.; Gentleman, E. Evolving insights in cell-matrix
interactions: Elucidating how non-soluble properties of the extrac-
ellular niche direct stem cell fate. Acta Biomater. 2015, 11, 3−16.
(43) Chung, T. W.; et al. Sialyllactose suppresses angiogenesis by
inhibiting VEGFR-2 activation, and tumor progression. ONCOTAR-
GET 2017, 35, 58152−58162.
(44) Mogaki, R.; Okuro, K.; Aida, T. Molecular glues for
manipulating enzymes: trypsin inhibition by benzamidine-conjugated
molecular glues. Chem. Sci. 2015, 6, 2802−2805.
6633
https://doi.org/10.1021/jacs.1c01951
J. Am. Chem. Soc. 2021, 143, 6622−6633