RESEARCH ARTICLE
www.mrc-journal.de

Identification of the First Sulfobetaine Hydrogel-Binding
Peptides via Phage Display Assay

Ramona B. J. Ihlenburg, David Petracek, Paul Schrank, Mehdi D. Davari,
Andreas Taubert,* and Dirk Rothenstein*

Dedicated to the memory of Prof. Wolfgang Meier, a great mentor, colleague, and friend

Using the M13 phage display, a series of 7- and 12-mer peptides which
interact with new sulfobetaine hydrogels are identified. Two peptides each
from the 7- and 12-mer peptide libraries bind to the new sulfobetaine
hydrogels with high affinity compared to the wild-type phage lacking a
dedicated hydrogel binding peptide. This is the first report of peptides binding
to zwitterionic sulfobetaine hydrogels and the study therefore opens up the
pathway toward new phage or peptide/hydrogel hybrids with high application
potential.

1. Introduction

Hydrogels are among the most intensely researched materials
these days. This is due to their advantageous and highly di-
verse properties like biocompatibility[1–4] or useful mechanical
properties.[5,6] Because of this, hydrogels have been explored for
numerous fields including medical applications (wound healing,
drug delivery, re-mineralization, etc.),[7,8] soft electronics,[8,9] or
as sensors and actuators.[10–12]

Zwitterionic hydrogels, a sub-class of the larger family of hy-
drogels, have attracted attention because of their rather unique

R. B. J. Ihlenburg, A. Taubert
Institute of Chemistry
University of Potsdam
Karl-Liebknecht-Straße 24–25, D-14476 Potsdam, Germany
E-mail: ataubert@uni-potsdam.de
D. Petracek, D. Rothenstein
Department Bioinspired Materials
Institute for Materials Science
University of Stuttgart
Heisenbergstraße 3, D-70569 Stuttgart, Germany
E-mail: dirk.rothenstein@imw.uni-stuttgart.de
P. Schrank, M. D. Davari
Department of Bioorganic Chemistry
Leibniz Institute of Plant Biochemistry
Weinberg 3, D-06120 Halle, Germany

The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/10.1002/marc.202200896

© 2023 The Authors. Macromolecular Rapid Communications published
by Wiley-VCH GmbH. This is an open access article under the terms of
the Creative Commons Attribution License, which permits use,
distribution and reproduction in any medium, provided the original work
is properly cited.

DOI: 10.1002/marc.202200896

properties and because of their appli-
cation potential in fields like medi-
cal materials,[13–15] electronics,[16–18]

sensing,[14,17,19] antifouling surfaces,[20,21]

or materials.[22,23] For example, Zhang et al.
have shown that poly(sulfobetaine) hydro-
gels are biocompatible, can be implanted,
and show properties and responses that are
comparable to poly(hydroxyethyl methacry-
late) hydrogels.[24] Similarly, He et al.[25]

showed that poly(sulfobetaine methacry-
late) hydrogels support wound healing
in mice. Specifically, the authors suggest

that softer and more viscoelastic hydrogels promote effects
that are beneficial to wound healing, such as cell proliferation,
granulation formation, collagen aggregation, and deposition of
chondrogenic extracellular matrix. Zhu et al.[26] made a dual-
functional zwitterionic hydrogel that serves as a glucose sensor,
pH sensor, and wound treatment for diabetic patients.

Besides medical applications, zwitterionic hydrogels have, for
example, found application as functional soft materials. Jiao et
al.[27] used proline zwitterions incorporated into hydrogels to
achieve highly stretchable, yet tough and transparent hydrogels
with high stability at temperatures down to −40 °C. They pos-
tulate that the resulting materials can be used for wearable elec-
tronics and stress sensing. Pei et al.[28] have shown a similar ap-
proach for a strain sensor for wireless monitoring of organ move-
ments. Liu et al.[29] demonstrated zwitterionic hydrogels with
superior antifouling and oil/water separation capabilities. Over-
all, these examples show the breadth of applications for zwitte-
rionic hydrogels and many more examples can be found in the
literature.[13–23,30,31]

We have previously studied hydrogel scaffolds for
biomaterials.[32–35] In this scope, we have developed a novel
sulfobetaine hydrogel based on the new dicationic cross-linker
N,N,N′,N′-tetramethyl-N,N′-bis(2-ethylmethacrylate)-propyl-
1,3-diammonium dibromide and the sulfobetaine monomer
2-(N-3-sulfopropyl-N,N-dimethyl ammonium)ethyl methacry-
late. The hydrogels and cryogels are effective for removing
organic dyes from aqueous solution[36] and have a rather
complex, multi-length scale hierarchical structure.[37] Overall,
hydrogel synthesis is straightforward and, like some of the exam-
ples introduced above, these hydrogels could be interesting for
application in membranes, antifouling surfaces, or biomaterials.
The unique aspect of these hydrogels is that there is a cationic
cross-linker, which is different from the more conventional

Macromol. Rapid Commun. 2023, 44, 2200896 2200896 (1 of 8) © 2023 The Authors. Macromolecular Rapid Communications published by Wiley-VCH GmbH



www.advancedsciencenews.com www.mrc-journal.de

hydrogels based on ethylene glycol cross-linkers. While these
traditional cross-linkers have been extensively studied, there
is essentially no information on hydrogels made from our
new cross-linker and the study therefore focuses on these new
hydrogels.

As relatively little is known about the preferences of these new
hydrogels toward biological or bioinspired entities, the current
study addresses the seemingly simple question of whether there
are certain peptide sequences that preferentially interact with
these sulfobetaine hydrogels cross-linked with the new, dicationic
cross-linker. Knowing these peptides and their binding character-
istics to the hydrogel may then 1) enable the construction of soft
hybrid materials (e.g., sensors or actuators), 2) allow for an opti-
mized design strategy yielding an improved interaction between
hydrogel and surrounding (biological) matrix (e.g., for biomate-
rials development), or 3) open up new pathways for phage-based
therapies.[38,39]

The phage display method, initially described in 1985,[40] is
a key technology to identify binding peptide sequences for dif-
ferent organic and inorganic materials. Nowadays, it is a high-
throughput screening method for protein–protein interactions
in molecular biology. In short, genetically manipulated bacterio-
phages (phages) display pseudo-randomized protein fragments
(peptides) as fusion proteins on the phage surface. The selec-
tion of peptides is based on the specific interaction of the pep-
tide/phage with a target substrate. The phage links the pheno-
type (peptide) and the genotype (genetic code) and thus the infor-
mation on the binding peptide can be easily extracted from the
phage DNA. M13 phages have a fiber-like proteinaceous capsid
of 900 nm length and 6 nm diameter that encapsulates a single-
stranded (ss) DNA genome, encoding five structural proteins and
six proteins for replication and assembly.[41] The major coat pro-
tein pVIII builds the phage body with ≈2700 copies that are heli-
cally arranged around the ssDNA. At the ends of the phage parti-
cle the minor coat proteins pIII and pVI at one end and pVII and
pIX at the other end are located.[42]

In recent years, the phage display method has been ex-
tended to investigate the interaction between peptides and in-
organic materials and thus, binding peptides for different mate-
rial classes, including semiconductors[43–45] and metals[46] have
been identified. The selected peptides have been applied in,
for example, biomineralization,[47] stress-resistant materials,[48]

piezo-active nanowires,[49] bio-sensing platforms,[50,51] and bat-
tery electrodes.[52,53]

Phage display technology is a method to select peptides based
on binding to a target molecule or surface. Therefore, phages are
genetically engineered to express an additional peptide fused to a
surface protein (coat protein) of the phage. Thereby, each phage
clone expresses one individual additional peptide sequence. In
the peptide library, 2.7 × 107 different phage clones (and thus the
same number of different peptides) are mixed. From this mix-
ture, the phage clones that bind to the target can be selected. The
phage has the task of physically connecting the phenotype with
the genotype, that is, that the binding peptide (phenotype) and
the corresponding DNA (genotype) can be isolated together. The
peptide sequence is then deduced from the corresponding DNA,
which encodes the peptide, by DNA sequencing.

Indeed, binding peptides directed against polymers have
been selected by phage display. The interaction was at-

tributed to the recognition of tacticity, that is, the position
of functional groups in the side chains, the structure, and
the crystallinity of the target system.[54] However, past re-
search mainly focused on hydrophobic polymer systems such
as poly(methyl methacrylate) (PMMA),[55,56] poly(l-lactide),[57]

polyetherimide,[58] polypropylene,[59] poly(dimethyl siloxane),[60]

and polystyrene/polyvinyl chloride.[61] A common sequence for
binding peptides has not been identified, but many of the bind-
ing peptides were enriched in aromatic and hydrophobic amino
acid residues.[62,63]

In contrast, only a few reports describe peptides bind-
ing to hydrophilic polymer systems, charged or uncharged.
Poly(propylene oxide), an uncharged polymer, strongly binds
to the peptide DFNPYLGVTPVL.[64] A few examples of pep-
tides interacting with hydrophilic and charged systems are
published. The peptide HNAYWHWPPSMT binds to poly(2-
methoxy-5-propyloxysulfonate-1,4-phenylenevinylene), and the
binding was suggested to be due to 𝜋–𝜋 interactions of tryp-
tophan residues.[65] Aromatic amino acids, mainly tryptophan,
in the peptide sequences are also beneficial for binding to
charged polymer nanoparticles.[66] In addition to aromatic amino
acids mediating the binding, peptides with charged amino acid
residues, mostly positively charged amino acids, bind to poly(N-
isopropylacrylamide).[67]

The current study, therefore, focuses on the identification
of peptides binding to our hydrogel based on the new cross-
linker.[36,37] Since, to the best of our knowledge, this work is
the first report on sulfobetaine hydrogel-binding peptides, we
will discuss similarities and differences with peptides interacting
with non-sulfobetaine hydrogels.[64,68] The study further expands
the pool of hydrogel-binding peptides to sulfobetaine hydrogels
and provides a hypothesis on the binding mechanism based on
computer analysis that opens avenues to next generation multi-
functional soft materials.

2. Results

2.1. Phage Display

Sulfobetaine hydrogel-binding peptides were selected from a ran-
dom 7- and 12-mer peptide library by phage display (Table 1 and
Tables S1 and S2, Supporting Information). Phage clones ex-
pressing the binding peptides were analyzed after the fourth and
fifth bio panning rounds. Eight different 12-mer peptides were
identified from 50 phage clones. The 7-mer peptides showed
greater diversity; here, 20 different peptide sequences were iso-
lated from 45 phage clones.

In the case of the 12-mer peptides (Table 1), the peptide
LLADTTHHRPWT clearly prevailed over other sequences with
40 out of 50 phage clones expressing this sequence. Besides this
sequence, seven other sequences were isolated. The phage clones
expressing 7-mer peptides were more diverse, with 20 different
peptide sequences identified. The most frequent sequence was
HGGVRLY with an occurrence of 8 out of 45. However, the se-
lection frequency of a peptide does not necessarily correlate with
the binding strength of a peptide.

The change in the abundance of individual amino acids in the
entirety of the selected peptides compared to the initial peptide li-
brary is an indicator of the importance of individual amino acids

Macromol. Rapid Commun. 2023, 44, 2200896 2200896 (2 of 8) © 2023 The Authors. Macromolecular Rapid Communications published by Wiley-VCH GmbH



www.advancedsciencenews.com www.mrc-journal.de

Table 1. Peptides binding to sulfobetaine hydrogels. 7- and 12-mer pep-
tide sequences which were applied in phage-based binding assays. In to-
tal, 20 different 7-mer peptide sequences from 45 phage clones and 8 dif-
ferent 12-mer peptide sequences from 50 clones were identified. Peptide
sequences are in single-letter amino acid code. Color code: red: positively
charged aa, blue: negatively charged aa, yellow: tryptophan. The peptide
sequences are sorted according to their length and calculated isoelectric
point (pI). The strongest binding peptides are printed bold.

Peptide # Sequence Length Frequency pI

18 QQTNWSL 7 4/45 5.50

107 QLAVAPS 7 4/45 5.50

56 GQSEKHL 7 6/45 7.19

26 KIAVIST 7 5/45 8.88

48 TVNFKLY 7 3/45 8.88

46 HGGVRLY 7 8/45 9.06

4 VPTFWTKIEHAA 12 1/50 7.19

7 LLADTTHHRPWT 12 40/50 7.38

80 NRPDSAQFWLHH 12 3/50 7.38

14 TMGFTAPRFPHY 12 1/50 9.06

for binding to the target substrate is (Table 2). The quotient of
the amino acid frequency in the hydrogel-binding peptides and
in the initial peptide library is indicative of enriched (criterion:
quotient > 1.25) or depleted (criterion: quotient < 0.75) amino
acid residues. Enriched amino acid residues may have a ben-
eficial effect on the binding of peptides, while depleted amino
acid residues might have an adverse effect on the interaction. For
both peptide lengths, 7- and 12-mer peptides, positively charged
amino acids were enriched while negatively charged amino acids
were depleted. Furthermore, aromatic amino acids with a hete-
rocyclic indole may also favor the peptide–hydrogel interaction.

A phage-based binding assay was performed to estimate the
binding strength of individual peptides (Figure 1). Since not all
clones could be tested, a selection was made (Table 2). The choice
of the 12-mer peptides included the two most abundant pep-
tides. In addition, two peptides were chosen that differed signif-
icantly in their isoelectric point. For the 7-mer peptides, the six
most frequently occurring peptides were selected. The binding
strength of the peptides was determined in parallel to wild-type
phages, that do not express a binding peptide, as negative con-
trol. This allows for an exclusion of possible contributions to the
interaction from the phage itself. Furthermore, no changes in the
peptide sequence or hydrogel composition has to be made. The

phage-based binding assay also showed no direct correlation be-
tween the frequency of isolation of a phage clone and the binding
strength of the peptide.

The abundance of a particular peptide sequence cannot be di-
rectly equated with the binding strength, since only a random
sample of the isolated phage clones was sequenced and biologi-
cal factors, such as the infectivity of a clone, may also influence
the abundance. The values of the binding strength were normal-
ized to M13 wt that did not express an additional binding peptide
(wt phage titer = 1). Four peptides showed a significantly higher
binding strength compared to the wt phage. The binding strength
of the 7-mer peptides 56 and 18 was increased by a factor of 3, and
doubled for the 12-mer peptides 4 and 80. Two 7-mer peptides
(107, 26) and the 12-mer peptide 7 showed a binding strength
similar to the wt phages. Only three peptides (7-mer: 46, 48; 12-
mer: 14) showed poorer binding compared to the wt phages.

To explain the experimentally observed trends in the binding
strengths, electrostatic potential map analyses were carried out
for all peptides. First, 3D models for the respective peptides were
generated using the AlphaFold2 software.[69] The generated pep-
tide models were relaxed using the AMBER14 force field[70] and
afterward scored using lDDT values.[71] The peptide models gen-
erally showed good lDDT values (see Table S3, Supporting Infor-
mation), indicating reliable peptide models, suitable for further
analyses.

The predicted structures show that the propensity of most pep-
tides to adopt a defined structure is very weak. This possibly sug-
gests that the binding strength correlates with the electrostatic
potential exhibited by the peptides. To confirm this, we generated
electrostatic potential surfaces for all peptide models using the
Adaptive Poisson–Boltzmann Solver (APBS).[72] To visually con-
tour the potential charges on the surface a three-color red (−1),
white (0), and blue (+1) scale APBS was rendered in PyMOL, col-
oring negatively, neutrally, and positively charged patches in red,
white, and blue, respectively (Figure 2 and Figure S1, Supporting
Information).

As a general trend, the electrostatic potentials showed that
most peptides, excluding 12-mer peptide 7, display a polarized
distribution of charges. One side of the peptide contains a posi-
tive patch and the other side contains a negative patch, separated
by a neutral zone of variable size (Figure 2A,B).

A closer look at the electrostatic potential surface for the
hydrogel-binding peptides showed that this charge distribution
could be also observed in almost all selected peptides (see Table
S3, Supporting Information). As a result, it is highly likely that

Table 2. Frequency of amino acid residues of hydrogel-binding peptides. The frequencies are given as quotients of the amino acid frequency in the
selected peptides, and the initial peptide library (NEB), that is, numbers >1 indicate enrichment and numbers <1 indicate depletion of the amino acid.
Each peptide sequence was considered only once for the calculation of amino acid abundance, regardless of its number of occurrences in the peptide
pool. The threshold value for enriched and depleted amino acids was set to 1.25 and 0.75, respectively. Enriched amino acid residues are highlighted in
dark grey and depleted amino acid residues are highlighted in light grey. Amino acids are indicated in one-letter code.

Peptide length Basic Small Hydrophobic Acidic Amide Nucleophilic Aromatic

H K R G A V L I P M D E N Q S T Ca) Y F W

7-mer 0.78 1.30 0.87 1.06 1.81 1.00 0.85 1.05 0.63 0.77 0.33 0.58 1.26 1.02 1.05 1.46 0.00 0.66 1.06 1.95

12-mer 1.98 1.49 1.48 2.14 2.08 0.71 1.64 0.41 1.48 1.07 0.99 0.44 0.91 0.54 0.97 1.25 0.00 0.77 1.68 3.16

a)
The abundance of cysteine residues was not considered because cysteine interferes with the infection of bacteria and phage formation, therefore the values may not be

based on binding.

Macromol. Rapid Commun. 2023, 44, 2200896 2200896 (3 of 8) © 2023 The Authors. Macromolecular Rapid Communications published by Wiley-VCH GmbH



www.advancedsciencenews.com www.mrc-journal.de

Figure 1. Relative binding strength of selected peptides to sulfobetaine hydrogels. The binding strength is normalized to the wild-type (wt) phage not
expressing an additional peptide. A) 7-mer peptides provide up to threefold higher binding strength. B) 12-mer peptides show up to twofold higher
binding strength to the hydrogel than the wt.

this polarization of the electrostatic potential is essential for the
effective binding of the peptide to the hydrogel through one side
without having repulsive effects from the opposite charge (see
Figures S1 and S2, Supporting Information). In the case of the
12-mer peptide 7, two distinct positive and negative patches were
flanked by one another (Figure 2B), suggesting repulsive effects
that affect the binding and lead to a decrease in the overall bind-
ing strength.

Second, the electrostatic potentials also displayed peptides con-
taining larger negative patches, correlating with higher binding
strength to the hydrogel overall. Coupled with this observation
(that peptides with pI at neutral to slightly acidic pH values also
show better binding strength than their basic counterparts) it can
be assumed that binding predominately happens through the
negatively charged regions of the peptides.

In addition, a positively charged peptide stretch is also neces-
sary for a strong interaction. Again, 12-mer peptide 7 is an out-
lier from this observation, as it contains two relatively large neg-
atively charged patches while exhibiting a relatively small bind-
ing strength of 1.2 (Figure 2B). This decreased binding strength
is to be expected, as the negatively charged binding regions are
flanked by positively charged regions, resulting in partially repul-
sive effects.

3. Discussion

To our best knowledge, sulfobetaine hydrogels have never been
studied in phage display experiments before. We succeeded in the
selection of peptides, which bind to the zwitterionic hydrogel and
identified some preliminary design principles of such zwitterion
hydrogel-binding peptides.

In general, the large majority of peptides contain charged
amino acids. The amino acids often occur in so-called dou-
blets, tandems of positively and negatively charged amino acids.
Such doublet formation often occurs in proteins that are re-
lated to biomineralization.[74] Doublets were also found in 7-
mer peptides. In addition, positively charged amino acids were
often found in the identified peptide sequences. As shown for
inorganic-binding peptides,[44] the amino acid histidine is also

significantly enriched in peptides that interact with the sulfobe-
taine hydrogels.

Beside charged amino acids, the abundance of the aromatic
amino acid residues of the selected peptides changes significantly
compared to the starting peptide library. While tyrosine and
phenylalanine were depleted in the selected hydrogel-binding
peptides, tryptophan was highly enriched and present in the ma-
jority of peptides with the best binding properties.

This is a rather interesting observation, as the hydrogel is
charged and has an overall positive charge stemming from the
cross-linker. We currently speculate (as there are no real possi-
bilities of 𝜋–𝜋 interactions in the hydrogel) that the strong en-
richment and consequently effective binding of tryptophan units
could be due to hydrogen bonding between the nitrogen atom
in the tryptophan side chain and the peptide (possibly, the water
molecules in the hydrogel are strongly involved in this process).
Molecular simulation-based materials informatics also suggest
that tryptophan plays a central role in the binding of peptides
to PMMA hydrogels.[75] The binding affinities (binding free en-
ergies) were obtained with molecular simulations based on the
response surface method (Kriging method). There, the calcu-
lated detachment energies of tryptophan from COOCH3 and
CH3 were the highest of the biogenic amino acids.

All peptides with good binding properties to the sulfobetaine
hydrogels (12-mer peptides 4, 80, 7, and 7-mer peptides 56, 18,
107, 26; Figure 2) have a calculated negative charge at pH 7.5 be-
tween−0.2 and−0.1 (in spite of the fact that most peptides are en-
riched in basic, i.e., positively charged amino acid residues). This
indicates that the main binding mode may be strongly driven by
electrostatics because the hydrogel has a positive overall charge.

The reason is as follows: the SPE monomer carries one posi-
tive and one negative charge, hence is neutral overall. However,
the cross-linker carries two positive charges, see refs. [36, 37]. As
stated in the experimental part, the hydrogel synthesis starts with
12 mmol of the SPE monomer and 1.2 mmol of the cross-linker,
yielding a 10:1 molar ratio of monomer:cross-linker. Assuming
full conversion (which is, however, highly unlikely), there would
be 12 mmol of ammonium groups and 12 mmol of sulfonate
groups (both from the SPE monomer) in the final hydrogel, and
this would effectively lead to a net charge neutralization.

Macromol. Rapid Commun. 2023, 44, 2200896 2200896 (4 of 8) © 2023 The Authors. Macromolecular Rapid Communications published by Wiley-VCH GmbH



www.advancedsciencenews.com www.mrc-journal.de

Figure 2. Comparison of electrostatic potential surfaces for models of 7-mer peptides 56, 18, 107, 26, 46, and 48, and 12-mer peptides 4, 80, 7, and 14. The
charge maps for representation of the electrostatic potential surfaces were calculated using the Adaptive Poisson–Boltzmann Solver (APBS) software[72]

as a Python package imported for PyMOL2 software.[] The distributions of the charges were mapped on the molecular surfaces of the peptides model
and contoured based on the APBS coloring scale rendered within PyMOL. Negatively charged surface patches are shown in red, neutrally charged surface
patches in white, and positively charged surface patches in blue. The surfaces are shown from two views of the same perspectives, related by a 180°

rotation around a vertical axis. A) Amino acid sequence, binding strength, isoelectric point, and electrostatic potential surfaces for the 7-mer peptides
56, 18, 107, 26, 46, and 48 ordered from top to bottom by decreasing binding strength. B) Amino acid sequence, binding strength, isoelectric point, and
electrostatic potential surfaces for the 12-mer peptides 4, 80, 7, and 14 ordered from top to bottom by decreasing binding strength. C) APBS coloring
scale for contouring electrostatic potential charges on the molecular surface of the peptide models. The scale consists of a three-color red (−1)—white
(0)—blue (+1) code for coloring of negative, neutral, and positive patches, respectively.

However, there is also 1.2 mmol of the cross-linker, and as
each cross-linker molecule carries two positive charges, there is
2.4 mmol of positively charged units in the hydrogel at full con-
version. This excess positive charge is very likely a key source for
the peptide/hydrogel interaction. In addition, hydrogen bonding,
dipolar interaction, and further ionic interactions between both
the ester groups in the polymer backbone and the peptide and be-
tween the charged entities of the SPE monomer and the peptides
are further sources of interaction.

It must, however, clearly be stated at this point that this charge
distribution can drastically change at different pH. Lower pH
values will significantly drive up the degree of protonation on
the basic amino acid residues in the peptides and, consequently,
also lead to a protonation of the carboxylate groups in the acidic
residues. As a result, there will be situations, where the patches as

shown in Figure 2 will not carry their specified charge anymore.
Negatively charged patches can turn neutral due to protonation of
the carboxylates and some of the unprotonated nitrogen atoms in
the basic residues can further be protonated. This will then poten-
tially lead to a repulsive interaction between a positively charged
cross-linker and a positively charged peptide residue. Based on
our findings and the visualization of the surface electrostatic po-
tential, we can conclude preliminary design principles for sulfo-
betaine hydrogel-binding peptides.

Three categories of binding strength (strong, medium, low)
were distinguished based on the molecular properties of the
peptides: 1) strong binders have adjacent positive and negative
stretches, which are not separated by other amino acid residues
(peptides 56, 18, 4, and 80). 2) Medium binders have the charged
peptide stretches separated by a stretch of neutral amino acids

Macromol. Rapid Commun. 2023, 44, 2200896 2200896 (5 of 8) © 2023 The Authors. Macromolecular Rapid Communications published by Wiley-VCH GmbH



www.advancedsciencenews.com www.mrc-journal.de

(peptides 107, 26, and 7). 3) Finally, peptides, showing limited
binding capacity only have a small positively charged region on
an otherwise negatively charged peptide surface (peptides 46, 48,
and 14).

Although we therefore postulate that one of the main driving
forces for peptide–hydrogel interaction is a negative electrostatic
potential surface, positive charges seem to be necessary for an ef-
fective peptide binding, too. Positively charged amino acids can
balance the charges of the bound peptide, resulting in a stable
binding. Moreover, positively charged amino acids may interact
with unreacted sulfobetaine residues of the hydrogel. Further-
more, peptides with flexible and large side chains in the negative
patches are better binders; this is visible from the computational
analyses. Larger side chains allow for more flexibility of the pep-
tide and thus a better occupancy of suitable binding sites in the
hydrogel. The superiority of a flexible peptide ligand may also re-
flect the structure of the hydrogel with fluctuations and disorder
on small size scales that can be overcome by flexibility.

4. Conclusion

The current study is the first binding assay of peptides to sulfobe-
taine hydrogels. The hydrogels are unusual because they do not
use a traditional, commercial cross-linker. The combination of
phage display and computational analysis clearly provides a first
glimpse into how rather short and essentially disordered peptides
interact with sulfobetaine hydrogels. Overall, the data show that
phage display analysis coupled with computational analysis can
1) be used to elucidate why and how certain peptides bind to sul-
fobetaine hydrogels and 2) provide inspirations for the design
of new materials or surfaces where the interaction can be tuned
(and possibly alternated between an ON and OFF state) by way
of the specific peptide sequence. Potential applications are phage-
reinforced hydrogels, hydrogels with specific sensing properties
provided by the phages, materials for affinity chromatography, or
materials for soft robotics that react to multiple stimuli. Further-
more, peptide/polymer hybrids have been suggested for separat-
ing biomacromolecules, where one specific peptide adheres to,
for example, a polymer hydrogel particle while others do not[66]

resulting in the separation of bio(macro)molecule by selective ad-
sorption for various analytical purposes.

5. Experimental Section
Materials: Yeast extract, tryptone, agar, EDTA, NaCl, polyethylene

glycol (PEG)-8000, PEG-8000/sodium chloride (PEG/NaCl), Tris, Tween
20, glycine, bovine serum albumin (BSA), HCl (37%), isopropyl-𝛽-D-
thiogalactopyranoside, 5-bromo-4-chloro-3-indolyl-𝛽-D-galactopyranoside
(Xgal), E.Z.N.A. M13 DNA Mini kit (Omega Bio-tek, USA), Monarch Plas-
mid miniprep kit (New England Biolabs, USA), and phage libraries (Ph.D.
-12 Phage Display Peptide Library and Ph.D. -7 Phage Display Peptide Li-
brary, New England Biolabs, USA), were purchased as p.a. quality from
Roth, Sigma-Aldrich, and Merck, and used without further purification.

The monomer 2-(dimethylamino ethyl)methacrylate (stabilized with
hydroquinone monomethyl ether for synthesis, Merck), 1,3-dibromo
propane (98%, Alfa Aesar), dimethyl formamide (water < 150 ppm, VWR
Prolabo), acetone (VWR, GPR RECTAPUR), tert-butyl methyl ether (99%,
Alfa Aesar), 2-(N-3-sulfopropyl-N,N-dimethyl ammonium) ethyl methacry-
late (SPE, Merck), and potassium peroxodisulfate (≥ 99%, Fluka Analyti-
cal) were used without further purification.

Hydrogel Synthesis: The cross-linker TMBEMPA/Br and the hydrogels
were synthesized as published.[37] In short, TMBEMPA/Br (1.2 mmol,
0.6196 g) and the monomer SPE (12 mmol, 3.3525 g) were dissolved in
7.2 mL of distilled water. This mixture was purged with N2 for 30 s. Then
KPDS (0.09 mmol, 24 mg) was added. Subsequently, 3 mL of this parent
solution were transferred to a 20 mL polypropylene container with a twist-
off-cap (SamcoTM Bio-TiteTM Specimen Containers, Thermo Scientific)
with an Eppendorf pipette. Before the containers were closed, the liquid
was purged again with N2 for 10 s. The containers were placed in a com-
partment drier (Memmert UF55Plus with grating, closed system setup,
ventilation 30%) at 70 °C for 20 min. The resulting, transparent to slightly
opaque gels were purified by immersion in 40 mL of ultrapure water (18.5
MΩ cm−1, 1 ppb TOC, 25 °C) for 4 weeks. The structure and properties of
the resulting gels were described previously.[37]

Phage Display M13 p3 Experiments: 7- and 12-mer peptides that specif-
ically bind to the hydrogels were isolated from random peptide libraries by
phage display (New England Biolabs, Inc.). The peptide libraries either ex-
press 7- or 12-mer peptides as N-terminal fusion protein with the minor
coat protein p3 of M13 phages with a short spacer sequence (GGGS) sepa-
rating the peptides from the p3 coat protein. General phage methods were
conducted according to the manufacturer’s recommendations.

The hydrogel binding substrates (≈5 × 5 × 5 mm3) were incubated at
least 1 h in Tris buffered saline (50 mm Tris pH 7.5; 150 mm NaCl) supple-
mented with 0.1% Tween 20 (TBST). The TBST was removed and 500 μL of
fresh TBST was added to the substrates. Subsequently, 10 μL of the phage
libraries (1 × 1013 pfu mL−1, pfu = plaque forming units) were added and
the mixtures were incubated at room temperature (RT) under constant ag-
itation (Thermomix, 800 rpm) for 60 min. Unbound phages were removed
from the substrate by washing ten times in TBS supplemented with 0.5%
Tween 20. The bound phages were eluted for 10 min at RT in 1 mL of
glycine elution buffer (0.2 m glycine, pH 2.2 supplemented with 1 mg mL−1

BSA). The eluate was neutralized with 150 μL of 1 m Tris-HCl, pH 9.1. The
eluted phages were amplified in Escherichia coli (E. coli) ER2738 in a shak-
ing incubator (250 rpm) for 4.5 h at 37 °C. The phages were purified by
polyethylene glycol-8000/sodium chloride (PEG/NaCl) precipitation. The
phage concentration after this first bio panning round was determined by
phage titering. Briefly, 200 μL of an E. coli culture in the exponential growth
phase were infected with 10 μL of the diluted phage suspension and plated
onto culture medium. The cultures were incubated overnight at 37 °C. The
pfu were counted and calculated on the concentration of the original phage
suspension.

In the following bio panning rounds, 1.5 × 1011 phages from the
previous bio panning round were applied. To increase the specificity of
hydrogel-binding peptides in total five panning rounds were performed.
From each experimental set up 30 randomly selected phage clones after
the fourth and fifth bio panning round were analyzed by DNA sequencing
(BioEdit Sequence Alignment Editor). The peptide properties, that is, iso-
electric point (pI) and the charge at pH 7.5, were calculated with Protein
Calculator v3.4.

Binding Assay M13 p3 Experiments: The binding strength of individual
hydrogel-binding peptides was determined with a phage-based binding as-
say. Phage clones expressing a hydrogel-binding peptide were amplified in
E. coli ER2738 and isolated by PEG/NaCl precipitation. Phages were re-
suspended in an appropriate volume of TBS and the concentration of the
phage stocks were determined by UV–vis spectroscopy and phage titering.

For the binding assay, a specific number of phages was incubated with
a defined volume of hydrogel in TBS at RT for 1 h. Unbound phages were
removed by washing with TBS supplemented with 0.5% Tween 20. Bound
phages were eluted with glycine elution buffer. The phage titer was deter-
mined by UV–vis spectroscopy and phage titering using appropriate dilu-
tions of the eluted phage suspensions. For the phage titering, only plates
with a minimum of 40 pfu were considered in the concentration calcula-
tions. Since the infection values were subject to fluctuations, the phage
titration was performed four times. The phage titer directly correlates with
the binding strength of the hydrogel-binding peptide. To facilitate com-
parison, the values were normalized to the wild type phages (wt phage
titer = n1).

Macromol. Rapid Commun. 2023, 44, 2200896 2200896 (6 of 8) © 2023 The Authors. Macromolecular Rapid Communications published by Wiley-VCH GmbH



www.advancedsciencenews.com www.mrc-journal.de

Molecular Modeling: Peptide structures were predicted using
AlphaFold2[69] by inputting the peptide sequences in FASTA formatting
and running the algorithm with default settings. The predicted structures
(a total of five for each peptide) were relaxed and energy minimized by
AlphaFold2 using the AMBER 14.0 force field.[70] The relaxed models were
then scored by Alphafold2 using lDTT[71] calculations and ranked accord-
ingly. The best ranked model for each peptide was used for electrostatics
calculation with the APBS using the PDB2PQR[76] and APBS solver[72]

software packages for Python imported in PyMOL2. Protonation states of
the amino acid residues were assigned based on PropKa calculations at
pH 7.[77] Generation and analysis of the electrostatic potential surfaces for
the peptides was performed using PyMOL2 (Schrödinger) software.[73]

For this, a custom PyMOL script was developed to automatize generation
and analysis called “peptide_analysis.pml.” The workflow is described
in the Method S1, Supporting Information. Detailed information on
the plugin, the code, and how to use it is found on GitHub under
“https://github.com/Protein-Engineering-Framework/APBS-peptides.”

Supporting Information
Supporting Information is available from the Wiley Online Library or from
the author.

Acknowledgements
The authors thank the University of Potsdam for financial support (A.T.,
grant 53170000). D.R. was in part supported by the Volkswagen Stiftung
under the project “Self-assembled nano-reactors via bacteriophage engi-
neering (ROBOPHAGE),” and the DFG (RO-3965/3-1). D.R. also thanks
the University of Stuttgart for support with laboratory infrastructure and
space and financial support. M.D.D. is supported through funds from the
Leibniz Institute of Plant Biochemistry.

Open access funding enabled and organized by Projekt DEAL.

Author Contributions
The manuscript was written through the contributions of all authors. All
authors have given approval to the final version of the manuscript.

Conflict of Interest
The authors declare no conflict of interest.

Data Availability Statement
The data that support the findings of this study are available from the cor-
responding author upon reasonable request.

Keywords
hydrogels, LLADTTHHRPWT, peptide binding, phage displays, QQT-
NWSL, sulfobetaine, surface modifications, TVNFKLY

Received: November 15, 2022
Revised: January 11, 2023

Published online: February 16, 2023

[1] E. Caló, V. V. Khutoryanskiy, Eur. Polym. J. 2015, 65, 252.

[2] J. Saroia, W. Yanen, Q. Wei, K. Zhang, T. Lu, B. Zhang, Bio-Des. Manuf.
2018, 1, 265.

[3] X. Xue, Y. Hu, Y. H. Deng, J. C. Su, Adv. Funct. Mater. 2021, 31,
2009432.

[4] L. L. Palmese, R. K. Thapa, M. O. Sullivan, K. L. Kiick, Curr. Opin.
Chem. Eng. 2019, 24, 143.

[5] S. Bashir, M. Hina, J. Iqbal, A. H. Rajpar, M. A. Mujtaba, N. A. Al-
ghamdi, S. Wageh, K. Ramesh, S. Ramesh, Polymers 2020, 12, 2702.

[6] M. Shibayama, Soft Matter 2012, 8, 8030.
[7] S. Correa, A. K. Grosskopf, H. L. Hernandez, D. Chan, A. C. Yu, L. M.

Stapleton, E. A. Appel, Chem. Rev. 2021, 121, 11385.
[8] J. H. Li, C. T. Wu, P. K. Chu, M. Gelinsky, Mater. Sci. Eng., R 2020, 140,

100543.
[9] H. Yuk, B. Lu, X. Zhao, Chem. Soc. Rev. 2019, 48, 1642.

[10] J. Shang, X. Le, J. Zhang, T. Chen, P. Theato, Polym. Chem. 2019, 10,
1036.

[11] Y. S. Zhang, A. Khademhosseini, Science 2017, 356, eaaf3627.
[12] X. Y. Liu, J. Liu, S. T. Lin, X. H. Zhao, Mater. Today 2020, 36, 102.
[13] J. Wu, Z. Xiao, A. Chen, H. He, C. He, X. Shuai, X. Li, S. Chen, Y.

Zhang, B. Ren, J. Zheng, J. Xiao, Acta Biomater. 2018, 71, 293.
[14] L. F. Wang, G. R. Gao, Y. Zhou, T. Xu, J. Chen, R. Wang, R. Zhang, J.

Fu, ACS Appl. Mater. Interfaces 2019, 11, 3506.
[15] X. Qiu, J. M. Zhang, L. L. Cao, Q. Jiao, J. H. Zhou, L. J. Yang, H. Zhang,

Y. P. Wei, ACS Appl. Mater. Interfaces 2021, 13, 7060.
[16] X. J. Sui, H. S. Guo, P. G. Chen, Y. N. Zhu, C. Y. Wen, Y. H. Gao, J.

Yang, X. Y. Zhang, L. Zhang, Adv. Funct. Mater. 2020, 30, 1907986.
[17] C. J. Lee, H. Y. Wu, Y. Hu, M. Young, H. F. Wang, D. Lynch, F. J. Xu, H.

B. Cong, G. Cheng, ACS Appl. Mater. Interfaces 2018, 10, 5845.
[18] S. E. Feicht, A. S. Khair, Soft Matter 2016, 12, 7028.
[19] W. Yang, H. Xue, L. R. Carr, J. Wang, S. Y. Jiang, Biosens. Bioelectron.

2011, 26, 2454.
[20] H. M. Chan, N. Erathodiyil, H. Wu, H. F. Lu, Y. R. Zheng, J. Ying, Mater.

Today Commun. 2020, 23, 100950.
[21] W. Pan, T. J. Wallin, J. Odent, M. C. Yip, B. Mosadegh, R. F. Shepherd,

E. P. Giannelis, J. Mater. Chem. B 2019, 7, 2855.
[22] L. D. Blackman, P. A. Gunatillake, P. Cass, K. E. S. Locock, Chem. Soc.

Rev. 2019, 48, 757.
[23] W. J. Yang, K.-G. Neoh, E.-T. Kang, S. L.-M. Teo, D. Rittschof, Prog.

Polym. Sci. 2014, 39, 1017.
[24] Z. Zhang, T. Chao, L. Liu, G. Cheng, B. D. Ratner, S. Jiang, J. Biomater.

Sci., Polym. Ed. 2009, 20, 1845.
[25] H. He, Z. Xiao, Y. Zhou, A. Chen, X. Xuan, Y. Li, X. Guo, J. Zheng, J.

Xiao, J. Wu, J. Mater. Chem. B 2019, 7, 1697.
[26] Y. Zhu, J. Zhang, J. Song, J. Yang, Z. Du, W. Zhao, H. Guo, C. Wen, Q.

Li, X. Sui, L. Zhang, Adv. Funct. Mater. 2020, 30, 1905493.
[27] Q. Jiao, L. Cao, Z. Zhao, H. Zhang, J. Li, Y. Wei, Biomacromolecules

2021, 22, 1220.
[28] X. Pei, H. Zhang, Y. Zhou, L. Zhou, J. Fu, Mater. Horiz. 2020, 7, 1872.
[29] H. Liu, L. Yang, B. Dou, J. Lan, J. Shang, S. Lin, Sep. Purif. Technol.

2021, 279, 119789.
[30] S. H. Liu, J. Y. Tang, F. Q. Ji, W. F. Lin, S. F. Chen, Gels 2022, 8, 46.
[31] N. Erathodiyil, H. M. Chan, H. Wu, J. Y. Ying, Mater. Today 2020, 38,

84.
[32] S. Prieto, A. Shkilnyy, C. Rumplasch, A. Ribeiro, F. J. Arias, J. C.

Rodríguez-Cabello, A. Taubert, Biomacromolecules 2011, 12, 1480.
[33] A. Shkilnyy, R. Gräf, B. Hiebl, A. T. Neffe, A. Friedrich, J. Hartmann,

A. Taubert, Macromol. Biosci. 2009, 9, 179.
[34] M. Schneider, C. Gunter, A. Taubert, Polymers 2018, 10, 275.
[35] T. Mai, K. Wolski, A. Puciul-Malinowska, A. Kopyshev, R. Graf, M.

Bruns, S. Zapotoczny, A. Taubert, Polymers 2018, 10, 1165.
[36] R. B. J. Ihlenburg, A. C. Lehnen, J. Koetz, A. Taubert, Polymers 2021,

13, 208.
[37] R. B. J. Ihlenburg, T. Mai, A. F. Thünemann, R. Baerenwald, K. Saal-

wächter, J. Koetz, A. Taubert, J. Phys. Chem. B 2021, 125, 3398.

Macromol. Rapid Commun. 2023, 44, 2200896 2200896 (7 of 8) © 2023 The Authors. Macromolecular Rapid Communications published by Wiley-VCH GmbH



www.advancedsciencenews.com www.mrc-journal.de

[38] H. Y. Kim, R. Y. K. Chang, S. Morales, H. K. Chan, Antibiotics 2021, 10,
130.

[39] I. M. Martins, R. L. Reis, H. S. Azevedo, ACS Chem. Biol. 2016, 11,
2962.

[40] G. P. Smith, Science 1985, 228, 1315.
[41] D. Marvin, Curr. Opin. Struct. Biol. 1998, 8, 150.
[42] G. Å. Løset, I. Sandlie, Methods 2012, 58, 40.
[43] S. Kilper, T. Jahnke, K. Wiegers, V. Grohe, Z. Burghard, J. Bill, D.

Rothenstein, Materials 2019, 12, 904.
[44] D. Rothenstein, B. Claasen, B. Omiecienski, P. Lammel, J. Bill, J. Am.

Chem. Soc. 2012, 134, 12547.
[45] D. Rothenstein, D. Shopova-Gospodinova, G. Bakradze, L. P. H. Jeur-

gens, J. Bill, CrystEngComm 2015, 17, 1783.
[46] H. Heinz, B. L. Farmer, R. B. Pandey, J. M. Slocik, S. S. Patnaik, R.

Pachter, R. R. Naik, J. Am. Chem. Soc. 2009, 131, 9704.
[47] D. Rothenstein, S. J. Facey, M. Ploss, P. Hans, M. Melcher, V. Srot,

P. A. v. Aken, B. Hauer, J. Bill, Bioinspired, Biomimetic Nanobiomater.
2013, 2, 173.

[48] S. Kilper, S. J. Facey, Z. Burghard, B. Hauer, D. Rothenstein, J. Bill,
Adv. Funct. Mater. 2018, 28, 1705842.

[49] S. Kilper, T. Jahnke, M. Aulich, Z. Burghard, D. Rothenstein, J. Bill,
Adv. Mater. 2019, 31, 1805597.

[50] M. Adhikari, S. Dhamane, A. E. V. Hagström, G. Garvey, W.-H. Chen,
K. Kourentzi, U. Strych, R. C. Willson, Analyst 2013, 138, 5584.

[51] M. Adhikari, U. Strych, J. Kim, H. Goux, S. Dhamane, M.-V. Poonga-
vanam, A. E. V. Hagström, K. Kourentzi, J. C. Conrad, R. C. Willson,
Anal. Chem. 2015, 87, 11660.

[52] Y. J. Lee, H. Yi, W.-J. Kim, K. Kang, D. S. Yun, M. S. Strano, G. Ceder,
A. M. Belcher, Science 2009, 324, 1051.

[53] K. T. Nam, R. Wartena, P. J. Yoo, F. W. Liau, Y. J. Lee, Y.-M. Chiang,
P. T. Hammond, A. M. Belcher, Proc. Natl. Acad. Sci. USA 2008, 105,
17227.

[54] K. A. Günay, H.-A. Klok, Bioconjugate Chem. 2015, 26, 2002.
[55] T. Serizawa, T. Sawada, H. Matsuno, Langmuir 2007, 23, 11127.
[56] J. Chen, T. Serizawa, M. Komiyama, Angew. Chem., Int. Ed. 2009, 48,

2917.
[57] H. Matsuno, J. Sekine, H. Yajima, T. Serizawa, Langmuir 2008, 24,

6399.
[58] T. Date, J. Sekine, H. Matsuno, T. Serizawa, ACS Appl. Mater. Interfaces

2011, 3, 351.
[59] C. Juds, J. Schmidt, M. G. Weller, T. Lange, U. Beck, T. Conrad, H. G.

Börner, J. Am. Chem. Soc. 2020, 142, 10624.

[60] S. Swaminathan, Y. Cui, RSC Adv. 2012, 2, 12724.
[61] N. B. Adey, A. H. Mataragnon, J. E. Rider, J. M. Carter, B. K. Kay, Gene

1995, 156, 27.
[62] M. H. Caparon, P. A. De Ciechi, C. S. Devine, P. O. Olins, S. C. Lee,

Mol. Diversity 1996, 1, 241.
[63] M. Vodnik, U. Zager, B. Strukelj, M. Lunder, Molecules 2011, 16, 790.
[64] T. Serizawa, H. Fukuta, T. Date, T. Sawada, Chem. Commun. 2016, 52,

2241.
[65] H. Ejima, H. Kikuchi, H. Matsuno, H. Yajima, T. Serizawa, Chem.

Mater. 2010, 22, 6032.
[66] S.-H. Lee, I. Moody, Z. Zeng, E. B. Fleischer, G. A. Weiss, K. J. Shea,

ACS Appl. Bio Mater. 2021, 4, 2704.
[67] S. Suzuki, T. Sawada, T. Ishizone, T. Serizawa, Chem. Commun. 2016,

52, 5670.
[68] A. J. Keefe, K. B. Caldwell, A. K. Nowinski, A. D. White, A. Thakkar, S.

Jiang, Biomaterials 2013, 34, 1871.
[69] J. Jumper, R. Evans, A. Pritzel, T. Green, M. Figurnov, O. Ronneberger,

K. Tunyasuvunakool, R. Bates, A. Žídek, A. Potapenko, A. Bridgland,
C. Meyer, S. A. A. Kohl, A. J. Ballard, A. Cowie, B. Romera-Paredes, S.
Nikolov, R. Jain, J. Adler, T. Back, S. Petersen, D. Reiman, E. Clancy,
M. Zielinski, M. Steinegger, M. Pacholska, T. Berghammer, S. Boden-
stein, D. Silver, O. Vinyals, et al., Nature 2021, 596, 583.

[70] D. Case, V. Babin, J. Berryman, R. Betz, Q. Cai, D. Cerutti, T.
Cheatham, III, T. Darden, R. Duke, H. Gohlke, AMBER 14; University
of California: San Francisco 2014, 1–826.

[71] V. Mariani, M. Biasini, A. Barbato, T. Schwede, Bioinformatics 2013,
29, 2722.

[72] E. Jurrus, D. Engel, K. Star, K. Monson, J. Brandi, L. E. Felberg, D. H.
Brookes, L. Wilson, J. Chen, K. Liles, M. Chun, P. Li, D. W. Gohara,
T. Dolinsky, R. Konecny, D. R. Koes, J. E. Nielsen, T. Head-Gordon,
W. Geng, R. Krasny, G.-W. Wei, M. J. Holst, J. A. Mccammon, N. A.
Baker, Protein Sci. 2018, 27, 112.

[73] W. L. DeLano, The PyMOL Molecular Graphics System, Palo Alto, CA,
USA 2018.

[74] M. L. Lemloh, K. Altintoprak, C. Wege, I. M. Weiss, D. Rothenstein,
Materials 2017, 10, 119.

[75] T. Iwasaki, M. Maruyama, T. Niwa, T. Sawada, T. Serizawa, Polym. J.
2021, 53, 1439.

[76] T. J. Dolinsky, J. E. Nielsen, J. A. Mccammon, N. A. Baker, Nucleic
Acids Res. 2004, 32, W665.

[77] M. H. M. Olsson, C. R. Søndergaard, M. Rostkowski, J. H. Jensen, J.
Chem. Theory Comput. 2011, 7, 525.

Macromol. Rapid Commun. 2023, 44, 2200896 2200896 (8 of 8) © 2023 The Authors. Macromolecular Rapid Communications published by Wiley-VCH GmbH