ARTICLES

Biomaterials functionalization using a

novel peptide that selectively binds to a
conducting polymer
ARCHIT B. SANGHVI1*, KILEY P-H MILLER4*, ANGELA M.BELCHER3 AND CHRISTINE E. SCHMIDT2
1

Department of Biomedical Engineering, and 2Department of Chemical Engineering, The University of Texas at Austin, Austin, Texas 78712-1084, USA
Department of Materials Science and Engineering and 4Department of Bioengineering, Massachusetts Institute of Technology, Cambridge, Massachusetts
02139-4307, USA
*These authors contributed equally to this work

e-mail: schmidt@che.utexas.edu
3

Published online: 15 May 2005; doi:10.1038/nmat1397

The goal in biomaterial surface modification is to retain a

materials bulk properties while modifying only its surface
to possess desired recognition and specificity. Here we
develop a unique strategy for surface functionalization
of an electrically conductive polymer, chlorine-doped
polypyrrole (PPyCl), which has been widely researched
for various electronic and biomedical applications.
An M13 bacteriophage library was used to screen 109
different 12-mer peptide inserts against PPyCl. A binding
phage (T59) was isolated, and its binding stability and
specificity to PPyCl was assessed using fluorescence
microscopy and titer count analysis. The relative binding
strength and mechanism of the corresponding 12-mer
peptide and its variants was studied using atomic force
microscopy and fluorescamine assays. Further, the T59
peptide was joined to a cell adhesive sequence and used
to promote cell attachment on PPyCl. This strategy can
be extended to immobilize a variety of molecules to PPyCl
for numerous applications. In addition, phage display
can be applied to other polymers to develop bioactive
materials without altering their bulk properties.

here have been many techniques used to modify the surfaces of

biomaterials, including protein adsorption and self-assembly1,
direct covalent surface modifications25, and synthesis of
novel graft-copolymers with desired functional groups6. Physical
adsorption and self-assembly, which do not require chemical
processing, are unfortunately dependent upon unpredictable, nonspecific interactions between the protein and the surface1. On the
other hand, to modify a surface by covalent modification, a reactive
group must be present on the polymer, which is not always the case
for many polymers (for example, polypyrrole (PPy), polylactic acid).
Several researchers have previously modified polymers that lack a
functional chemical group by: (i) creating blends of the polymer
with another material that does possess covalent attachment
sites7,8, or (ii) synthesizing new polymers with reactive groups6.
Here we present a different approach for surface modification and
functionalization of synthetic polymers.
In the studies presented here, phage display was used to select
for peptides that specifically bind to an existing biomaterial, PPy,
and which can subsequently be used to modify the surface of PPy.
Previously, phage display has been used in various applications such
as identifying cancer-specific ligands9, receptorligand interactions10,
and in selecting unique peptides against inorganic semiconductor
materials1113. Reviews14,15 have highlighted the application of phage
display in selecting peptides to functionalize biomaterials such
as titanium. Other researchers have used phage display to select
peptides against synthetic polymers such as polystyrene16, which
was inadvertent, and yohimbine-imprinted methacrylate polymer
for molecular-imprinted receptors17. Although the use of phage
display has been widely studied, there have been few studies with
polymers, and even fewer investigations into using this approach
to functionalize polymeric scaffolds for specific biomedical
applications. This prompted our research to explore the versatility of
phage display in developing a method for surface functionalization
of a model biopolymer, chlorine-doped polypyrrole (PPyCl).
PPy is a conductive synthetic polymer that has numerous
applications in fields such as drug delivery18 and nerve

496

nature materials | VOL 4 | JUNE 2005 | www.nature.com/naturematerials

2005 Nature Publishing Group

nmat1397-print.indd 496

2005 Nature Publishing Group

9/5/05 4:47:15 pm

ARTICLES
a

a
Number of phage clones
bound to PPyCl (PFU)

Number of phage clones

(PFU l 1)

1010

107

104

10,000
1,000
100
10
1

T59

C511

S36
Phage

T125-4

WT

101
T59

T43

T36

WT

Phage

10,000
PPyCl

Number of phage clones

bound (PFU)

Number of phage clones

bound to PPyCl (PFU)

10,000
1,000
100
10
1
T59

T43

T36

Polypropylene

1,000

Polystyrene
100

10

WT

T59

Phage

Phage

C511

WT

Phage

12-mer peptide insert sequence

T59

Thr

His

Arg

Thr

Ser

Thr

Leu

Asp

Tyr

Phe

Val

Ile

T43

Ser

His

Lys

Tyr

Pro

Lys

Pro

Pro

Tyr

Phe

His

Trp

T36

Thr

Ile

Lys

Met

His

Thr

Leu

Ser

Tyr

Thr

Gly

Leu

Figure 1 Phage amplification and competitive binding demonstrating that

T59 was selected because of its specificity of binding PPyCl. a, Amplification
of the three clones acquired in the 5th biopan round (T36, T43, and T59) and
WT illustrates that they all grow at equal rates and that the final selection of
T59 was not influenced by a difference in growth. b, Competitive binding study
demonstrates that T59 binds significantly better by two orders of magnitude to
PPyCl when compared to both T36 and T43. The input concentration in b was
106 PFU per sample for each phage and n = 3 for both a and b. Note: The data
presented in b are on a log scale to better distinguish the binding of the clones.
Error bars represent the standard deviation.

regeneration19,20, and has been used in biosensors and coatings

for neural probes21,22. Different dopant ions such as chloride,
perchlorate, iodine and poly(styrene sulphonate) can be used during
electrochemical synthesis to provide the material with varying
properties (for example, conductivity, film thickness and surface
topography)2325. Although dopants that are relevant to biomedical
applications such as hyaluronic acid can be used to develop PPybased hybrid materials26, the choice of dopants is limited to the size
and charge (negative) of molecules. Furthermore, PPyCl does not
contain a functional group for biomolecule immobilization, making
it a suitable model polymer for functionalization using a peptide
selected with phage display.
To select peptides for PPyCl using phage display, we synthesized
thin films (200250 nm) of PPyCl by standard electrochemical
methods27,28. PPyCl thin films were incubated with a commercially
available 12-mer linear phage display library (M13 clone, filamentous
bacteriophage), which expressed random peptides on minor coat
proteins (pIII) with a diversity of 2.7 109 (Ph.D.-12 Phage Display
Peptide Library Kit, New England Biolabs)11,29,30. The sequences,
which correspond to the peptides displayed on the phage, were

Phage

12-mer peptide insert sequence

T59

Thr

His

Arg

Thr

Ser

Thr

Leu

Asp

Tyr

Phe

Val

Ile

C511

Met

Pro

Ala

Val

Met

Ser

Ser

Ala

Gln

Val

Pro

Arg

S36

Trp

Thr

Val

Gln

Thr

Pro

Thr

Met

Leu

Pro

Met

Met

T125-4

Asp

Ile

Ile

Ser

Ser

Trp

Tyr

Met

Pro

Asp

His

Pro

Figure 2 Titer count analysis of T59 binding to PPyCl compared with the
binding of other phage clones and the binding to various substrates. a, The
T59 shows significantly better binding to PPyCl compared with other phage
clones (T125-4 and C511, selected from other screens of PPyCl, and S36, an
inorganic semiconductor binding phage), and b, T59 exhibits insignificant binding
to other polymers (polypropylene (PP) and polystyrene (PS)). For T59, C511 and
WT, n = 6, and for S36 and T125-4, n = 3. Notes: The binding of C511 and
WT to PP and PS surfaces are too low to be distinguished on plot b. The data are
presented on a log scale to better distinguish the binding of the clones. Error bars
represent the standard deviation.

analysed using DNA sequencing, as described in ref. 11. Because

of the higher possibility of non-specific binding in the first two
rounds, DNA analysis was only performed on the last three rounds
of panning where the results yielded three different phage clones
(see Supplementary Information, Table SI). One phage (labelled
T59) with the 12-mer amino acid sequence (THRTSTLDYFVI)
was identified for further analysis because it appeared in 23 out of
25 phage clones analysed in the 5th biopan round. The probability of
randomly recovering this exact phage from the library is nearly zero (P
~ 4.485 106), as calculated from the percentage occurrence of each
amino acid illustrated in Table SII (Supplementary Information)29.
To ensure that the selection of T59 was not the result of
differences in its growth rate compared with other clones, an
amplification study was performed in which each clone was
incubated in a bacterial culture and amplified overnight. The results
in Fig. 1a demonstrate that the T59 growth rate is equal to that
497

nature materials | VOL 4 | JUNE 2005 | www.nature.com/naturematerials

2005 Nature Publishing Group

nmat1397-print.indd 497

2005 Nature Publishing Group

9/5/05 4:47:22 pm

ARTICLES
a

10 m

Figure 3 Qualitative assessment of T59 and T59 peptide binding to PPyCl. a, T59 (left) and C511 (right) binding to PPyCl was analysed using an antibodybiotin
streptavidinFITC label and fluorescence microscopy. Lack of fluorescence staining indicates that no binding of C511 to PPyCl was observed, suggesting that T59 is
selective to PPyCl. Scale bar, 100 m. b, Additionally, T59 peptide binding to PPyCl was visually assessed (left). The results indicate the preservation of T59 peptide binding
without expression through the phage. T59 peptides did not bind to polystyrene (middle) and polystyrenesulphonatedoped PPy (right). All samples were incubated with
equal concentrations of peptide (15 M) and streptavidinFITC. Scale bar, 100 m. c, SEM image of PPyCl illustrating the heterogeneity of the surface, which is consistent
with the morphology of T59 and T59 peptide binding observed in a and b.

for the other clones and wild-type phage, suggesting that T59 was
not selected in the biopanning process because of a difference in
growth rate. Additionally, a competitive binding study of the three
clones selected in the 5th and final biopan (T59, T36, T43)
round was performed (Table SI, Supplementary Information). The
results (Fig. 1b) illustrate the dominant binding of T59 to PPyCl by
two orders of magnitude with equal input concentrations (1 106
plaque forming units (PFU) per sample) of each phage. This further
confirmed the results of selection in which the occurrence of T59
was greater than ninety percent in the last biopan round.
To assess the binding specificity of T59 to PPyCl, T59 binding
was evaluated and compared with other phage clones (which
have different 12-mer sequences) and to other selected polymer
substrates. Phage binding was investigated using two methods:
titer count analysis and immunofluorescence microscopy. The use
of titer count analysis is semiquantitative and provided a binding
comparison of phage counts for T59, S36 (semiconductor
binding phage isolated in A. M. B.s laboratory), C511 and T125-4
(other selected phage from previous screens against PPyCl), and
wild-type (WT, non-engineered) phage. C511 and T125-4 were
selected from PPyCl in other screens using different 12-mer Ph.D.
12 peptide libraries. C511 is the 11th clone from the 5th biopan
round in one library screen, whereas T125-4 is the 4th clone from

the 5th round of a second library screen. The detailed results from
both of these screens are not presented in this study, because T59
clone was superior in binding to PPyCl and thus better suited for
functionalization applications.
Initial amounts of 1 106 PFU per sample of each phage were
incubated with PPyCl. Results from the binding study demonstrate
that T59 binds significantly better to PPyCl when compared
with other phage clones (Fig. 2a). Substrate binding specificity of
T59 to PPyCl was evaluated by comparing binding with other
polymers (polystyrene and polypropylene) (Fig. 2b). Polystyrene
and polypropylene were selected because they are common
polymers and are also part of the plasticware used in the biopanning
experiments. To visualize the extent of surface coverage of T59 on
PPyCl, the phage was labelled with a biotinylated anti-pIII protein
antibody and a secondary antibody, streptavidin-FITC, and imaged
with fluorescence microscopy (Fig. 3a). The results in Figs 2b and
3a demonstrate that T59 binds to PPyCl, but exhibits negligible
binding to other substrates, suggesting that the 12-mer sequence
expressed on the phage exhibits selective binding.
The 12-mer peptide sequence of T59 was independently
synthesized to evaluate the binding without the presence of the
phage, as the eventual application is for surface functionalization
of PPyCl using T59 peptide as a bifunctional linker. By adding a

498

nature materials | VOL 4 | JUNE 2005 | www.nature.com/naturematerials

2005 Nature Publishing Group

nmat1397-print.indd 498

2005 Nature Publishing Group

9/5/05 4:47:24 pm

ARTICLES
a
Sequence

T59

THRTSTLDYFVI

RT59

IVFYDLTSTRHT

FT59

THRTSTGGGGGG
GGGGGGLDYFVI

BT59
T59D

THRTSTLGYFVI

BRT59

GGGGGGIVFYDL

Surface binding

Rupture force (pN)

Variant

Approach (setpoint)

Retraction
Unbinding force ~112 39 pN
Deflection (nm)

% bound
(relative to T59)

100

*
*

Unbinding force (pN)

**
**

**
50

309

400

112 **

200

102 **

43 **
0

0
T59

RT59

FT59
BT59
T59 peptide variants

BRT59

GRGDS
PPyCl

T59D

T59
PPyCl

T59GRGDS
PPyCl

Biotin
streptavidin

Binding partners

Figure 4 Evaluation of binding of T59 peptide and its variants to PPyCl shows
the involvement of specific amino acids. a,b, To determine the role of critical amino
acids in T59 binding to PPyCl, an initial set of T59 peptide variants were synthesized
(a) and their binding was assessed with 15 M input concentration with a film size of
0.05 cm2 (b). The results were evaluated relative to T59 binding and demonstrate the
importance of aspartic acid (Asp, D) in T59 peptide binding to PPyCl. When aspartic
acid is present (T59, RT59, BT59, BRT59) binding is maintained, but diminishes with
its absence (FT59 and T59D). (*): T59 and RT59 binding is statistically significant
(p<0.05). (**): RT59, BT59, and BRT59 binding is statistically insignificant. For all
samples n = 6. Error bars represent the standard deviation.

Figure 5 Strength of T59 peptide binding to PPyCl. AFM tips (Au-coated Si2N4)
were functionalized with biotinylated BSA linked to streptavidin, which was used
to immobilize biotinylated-T59 peptide. a, Typical force plot: as the tip approaches
the surface, binding between T59 peptides and PPyCl results in a force of adhesion;
as the tip is retracted from the surface, force is required to rupture the bonds. b,
Mean unbinding force for T59PPyCl is compared to that of biotinstreptavidin
(n = 300 each). Modification of T59 with GRGDS does not alter its binding to PPyCl.
The strength of T59 and T59-RGD binding to PPyCl is statistically higher than that
of non-specifically adsorbed RGD to PPyCl (**p<0.05). Error bars represent the
standard deviation.

biotin group at the C-terminus (T59-GGGS-K-biotin), qualitative

binding was demonstrated using streptavidin-FITC labelling
(Fig. 3b). The lack of T59 peptide binding to both polystyrene and
polystyrene-sulphonate-doped PPy (PPyPSS) shows that the peptide
is inert to other surfaces, thereby excluding binding through nonspecific adsorption. When comparing T59 peptide binding to PPyCl
versus binding to PPyPSS, we can speculate that the binding is dopant
specific, but additional studies using dopant and polymer analogues
are necessary to elucidate the precise mechanism of binding and the
effects of structural topography. Additionally, T59 peptide did not
completely and uniformly cover PPyCls surface, as illustrated by the
fractured pattern (Fig. 3b). The reasons for this are not completely
understood, although the inherent heterogeneous surface topography
of PPyCl (illustrated in the electron microscope image in Fig. 3c and
also as previously demonstrated25) may be an important factor for the
actual binding of T59 and T59 peptides to PPyCl. Previous research
with polypyrrole films has established that anions have a significant
influence on the surface morphology25. This heterogeneity in surface
topography will be an important factor in the evaluation of PPyClspecific phage/peptide binding and surface coverage.
To quantify the surface coverage of T59 peptide and its variants
to PPyCl, a fluorescamine assay, which is commonly used to detect
the primary amino groups contained in amino acids, peptides, and
proteins31,32, was used to determine the bound concentration of T59
peptide or its variants at pH 7.4. By varying the input concentration
of T59 peptide, a saturation point (hundred percent coverage) was

determined on a controlled PPyCl sample size (two-sided surface,

0.5 cm2). The input concentration at which possible complete surface
coverage was reached was around 15 M (data not shown); these
parameters were used to evaluate the studies described below as it
was inconclusive exactly at which point saturation occurs. Future
work will consist of using better quantitative methods to elucidate
the saturation effect of T59 peptide binding to PPyCl.
To study the role of specific amino acids of T59 in binding
PPyCl, an initial set of peptide variants (Fig. 4a and Supplementary
Information, Table SIII) was synthesized. Initial targeted residues
of the T59 peptide were the acidic (carboxylic acid of aspartic acid
(D), position 8) and basic (imidazol of histidine (H) and guanidine
of arginine (R), position 2 and 3, respectively) functionalities based
on their physicochemical properties33. A fluorescamine assay was
used to compare the PPyCl-binding capability of variants in which
the charged amino acids (that is, HRD) were replaced and shuffled
(Fig. 4b). The results demonstrate that aspartic acid is a critical
amino acid for T59 peptide binding, where the binding of FT59
and T59D variants, which did not contain an aspartic acid residue,
was nearly zero. Studies with varying pH (data not shown) further
confirmed the significance of aspartic acid in modulating the binding
of T59 peptides to PPyCl; as the pH became more acidic (aspartic
acid R-group pKa ~ 4.0) and dropped below the pKa of Asp R-group,
the binding diminished significantly. From the lack of FT59 variant
binding, it is apparent that the N-terminal region is not critical in
binding PPyCl. In the Ph.D.12 M13 library used to select the T59,
499

nature materials | VOL 4 | JUNE 2005 | www.nature.com/naturematerials

2005 Nature Publishing Group

nmat1397-print.indd 499

2005 Nature Publishing Group

9/5/05 4:47:26 pm

ARTICLES
a

b
% cell adhesion

100

50

P-serum

TCPSserum

Pno serum

P-T59

PPP-GRGDS
T59GRGDS T59GRDGS

Sample condition

c
% bound
(relative to initial)

100

Serum medium

50

10 mM PBS (control)
0
0

10

14

16

18

21

Incubation time (days)

Figure 6 Functionalization of PPyCl using T59 peptide complexed with GRGDS

and T59 peptide binding stability to PPyCl. a, PC12 cells adhered (serum-free
medium) and are viable (after 48 h) when RGD is presented by T59 on PPyCl
(illustrated on the left using live/dead stain), and cells did not attach and survive
when T59 was present alone (right). Scale bar, 100 m. b, T59GRGDS promotes
PC12 cell attachment on PPyCl (P refers to PPyCl) under serum-free medium
conditions comparable to positive controls (tissue culture polystyrene and PPyCl
in serum-containing medium). PC12 cells did not adhere to PPyCl in the absence
of GRGDS presented by T59 peptide, as indicated by minimal cell adhesion
(P-no serum, P-T59, P-T59GRDGS, and P-GRGDS). Sample sizes were constant
(0.75 0.75 cm2) and where necessary 60 M peptide was incubated before cell
seeding at a density of 104 cells per sample. After 48 h, samples were analysed
using an MTS assay and binding was relative to TCPS-positive control set to 100
percent binding (n = 4). (*): the difference in percent cell adhesion between the
experimental (P-T59GRGDS) and positive controls is not statistically significant.
c, PPyCl-bound T59 peptides are relatively stable in the presence of serum proteins
for up to 21 days. Peptide binding stabilizes to 85% after 10 days of incubation in
serum-containing medium, indicating a strong and stable interaction (n = 6). Error
bars represent the standard deviation.

2.8% occurrence of aspartic acid was reported30 (highlighted in the

Supplementary Information, Table SII), indicating that other peptide
inserts in the library did contain aspartic acid residues. Additionally,
aspartic acid is present at positions 1 and 10 in clone T125-4
(Fig. 2). Based on this we speculate that although the presence of
aspartic acid in T59 peptide is critical for PPyCl recognition, it is

not the only essential amino acid (as T125-4, which contained
Asp residues in the peptide insert sequence, did not bind to PPyCl).
Additionally, when the aspartic acid residue is randomized near the
C-terminal end (BRT59), the binding is maintained, suggesting that
the actual position of aspartic acid is not essential. Furthermore, the
binding of variant RT59 (reverse T59) to PPyCl was not dramatically
reduced or eliminated when compared with T59, suggesting that
the sequence order of amino acids is not critical. One hypothesis
that could be extrapolated from these results is that the binding
is not sequence specific, but rather is composition specific. The
N-terminal end of the T59 sequence has a hydrophilic character,
suggesting its importance in solubility. Whereas the C-terminal end
contains the aspartic acid residue, which provides an acidic ion that
probably controls the binding to overall positively charged PPyCl;
a hydrophobic region surrounding the aspartic acid could provide
an affinity of its own for controlling the folding and binding of the
peptide. In the future, this hypothesis will be examined further using
both a combinatorial approach to replace specific amino acids (for
example, Asp replaced with Glu) and computational modelling to
elucidate a precise mechanism of interaction.
The relative strength of interaction between T59 peptides and
PPyCl was assessed using atomic force microscopy (AFM). AFM
has been used extensively for measuring picoNewton molecular
binding forces for streptavidinbiotin, antigenantibody and
peptideporphyrin molecules3436. As illustrated (Fig. 5), the
binding strength of T59 peptide to PPyCl was compared with
other biomolecular interactions such as the biotinstreptavidin
bond and non-specific adsorption. Based on earlier results from
the peptide variants, the T59-PPyCl interaction does not seem to
be sequence-specific, but rather is dependent on the composition
of amino acids. Thus, this particular AFM study was performed to
illustrate the relative strength of T59PPyCl binding, within the
range of strong non-covalent biotinstreptavidin forces to weak
antigenantibody interactions3436.
AFM studies were performed by modifying the AFM tip with
streptavidin and biotin, and then with biotinylated peptide. The
retract trace of the T59 peptide-modified AFM tip interacting with
the PPyCl surface exhibits an unbinding force before it reaches the set
point (Fig. 5a). The rupture force (adhesion force) was determined
for each binding partner using a constant loading rate (f = s c)
6,000 1,500 pN s1, where s is the spring constant and c is the
retraction velocity37. The binding strengths of T59 and T59-GRGDS
peptides to PPyCl were evaluated and compared to a well-known
streptavidinbiotin interaction34 (Fig. 5b). GRGDS, a cell adhesion
promoting peptide derived from extracellular matrix proteins
such as fibronectin3,4, was used for PPyCl functionalization by T59
peptide, as described below. Appropriate controls were performed to
ensure that the measurements correspond to the specific T59PPyCl
interactions and not to other nonspecific sources. These initial AFM
experiments did not selectively control for the number of T59 peptides
interacting with PPyCl. However, when compared with the controls
performed under the same conditions and using our same AFM
measurement techniques, the results suggest that T59 peptides bind
to PPyCl with relatively significant strength. Additional comparison
to the published and well-characterized binding strength between
a single streptavidinbiotin bond (highly specific and strongest
non-covalent bond known)34 results in the conclusion that T59
PPyCl binding has a minimal degree of specificity, as suggested by
earlier results with the peptide variants. When compared with nonspecific adsorption binding of GRGDS to PPyCl (Fig. 5b), T59 and
T59GRGDS interactions to PPyCl are statistically higher. Thus, the
strength of T59 binding to PPyCl is in the range of weak antigen
antibody forces35, and greater than published data for van der Waals
interactions (12 pN) and hydrogen bonding (17 pN)34,38. Future
work will be composed of detailed studies using varying loading

500

nature materials | VOL 4 | JUNE 2005 | www.nature.com/naturematerials

2005 Nature Publishing Group

nmat1397-print.indd 500

2005 Nature Publishing Group

9/5/05 4:47:28 pm

ARTICLES
rates37 and peptide concentrations to understand the various energy
barriers traversed in bond dissociation39 and to determine an actual
dissociation constant.
A model peptide (GRGDS) that promotes cell adhesion3,4,40,41
was used to examine the ability of the T59 peptide to promote
cell attachment and thus, to serve as a bifunctional linker. It has
been noted that to promote strong cell adhesion, stable linking of
biomolecules (for example, RGD) are necessary42. T59 peptide was
modified at the C-terminus to maintain consistency with the previous
biotin modification used for the qualitative peptide binding studies.
The activity of T59-GRGDS in promoting cell adhesion on PPyCl
was evaluated using PC12 cells, a rat pheochromocytoma cell line
that has been previously studied with polypyrrole19. PC12 cells do
not adhere to PPyCl in serum-free medium, but when modified with
T59-GRGDS peptides, PPyCl promoted cell adhesion, as illustrated
using a fluorescent live/dead stain (Fig. 6a). Cell quantification
using the MTS cell proliferation assay43 was also performed to
determine the extent of adhesion when compared with positive
controls (that is, cells adhered in the presence of serum to tissue
culture polystyrene and PPyCl), illustrated in Fig. 6b. The results
show that cells adhered to PPyCl modified with T59-GRGDS to the
same extent as positive controls, but did not adhere signigicantly
to PPyCl when modified with RGD non-specifically adsorbed (PGRGDS) and scrambled peptide (GRDGS) that is known to inhibit
cell adhesion42 (P-T59GRDGS, negative control). This supports our
hypothesis that the T59 peptide, selected by phage display, can be used
as a tool for biomolecule immobilization and to control biological and
cellular activity on PPyCl. Moreover, once the mechanism of T59 peptide
binding to PPyCl is completely elucidated, the T59 peptide can be further
tailored, if necessary, using rational-design-based approaches44 by either
altering non-essential amino acids or by adding spacer amino acids to
tether other biomolecules.
It is inherent in the design of biomaterials that the ultimate
application will be in the human body, and for that reason, it
is important to evaluate the stability of the T59 peptide under
physiological conditions where proteins (for example, albumin)
can non-specifically adsorb to the polymer surfaces45. Results
from fluorescamine studies, presented in Fig. 6c, illustrate that
the T59 peptide remains bound to PPyCl even after 21 days of
incubation in serum-containing medium at 37 C, suggesting a
stable interaction between the T59 peptide and the PPyCl surface.
The T59 peptide unbinding in serum-containing medium
stabilized after 10 days with an approximate loss of 15 7%.
When compared with control PBS conditions, where stabilization
was achieved after three days with a loss of 7 2%, the serum
condition had twice the loss. This indicates that serum proteins
can impact the unbinding of T59 peptides from PPyCl, but the
unbinding was non-specific and insignificant as stabilization was
reached with 85 10% of the original peptides still remaining
bound. The strength of T59-PPyCl interaction will be explored
quantitatively in future studies, along with the in vivo evaluation
of stability and biocompatibility.
The use of the selected peptide for PPyCl by phage display can
be extended to encompass a variety of therapies and devices such
as PPy-based drug delivery vehicles18, nerve guidance channel
conduits19,20, and coatings for neural probes22. Furthermore, this
strategy for surface functionalization could also serve as a versatile
method to develop bioactive hybrid-materials using other existing
polymers (including those that are already approved by the Federal
Drugs Administration and/or those polymers that lack functional
chemical groups for coupling reactions, like PPyCl), without
changing the bulk properties of the materials. Selection of peptides
using phage display thus represents a simple alternative to methods
based on electrostatic and hydrophobic interactions between two
moieties to achieve adsorptive modification of surfaces.

METHODS
More details on certain aspects of the experimental procedures are provided in the Supplementary
Information.
PPyCl films were electrochemically synthesized27 onto indium tin-oxide (ITO) conductive
borosilicate glass (Delta Technologies, Stillwater, Minnesota). Film thickness was determined by
integrating current over time and was controlled by passage of charge27 based on the standard value
of 50 mC cm2.
Peptides were selected for binding to PPyCl using a commercially available Ph.D.-12 library (New
England Biolabs, Beverly, Massachusetts), which provided 109 different phage clones with 12-amino acid
linear-peptide inserts11,29. After each round of selection and washing, the tightly bound phage clones
were eluted with low pH and amplified using bacterial medium (Escherichia coli strain ER2738, New
England Biolabs)11. A total of five cycles or rounds of screening were performed. The peptide sequences
of phage-expressed peptides specific to PPyCl were determined by DNA sequencing from The Institute
for Cellular and Molecular Biology (ICMB, The University of Texas, Austin, Texas).
Quantitative titer counts were used for amplification rate and competitive binding analysis (T59,
T43, T36) and comparison of binding to PPyCl between other clones (C511, T125-4, S36)
and substrates (for example, polystyrene (PS), polypropylene (PP)). For the growth rate study, T59,
T43, T36 and WT phage were separately amplified in bacterial culture medium for 5 h at 37 C and
their final count after 6 h of incubation was quantified. The phage solutions were titered to obtain
amplification of the clones. A competitive binding study with the same phage was also performed (see
Supplementary Information). For phage (T59 compared to C511, T125-4, S36) and substrate
(PPyCl versus PS, PP) binding comparison, the specific phage was incubated (106 PFU per sample) in
substrate-containing tubes under the same conditions as described for peptide selection. Binding was
assessed by counting PFU (see Ph.D. 12 manual from New England Biolabs).
Qualitative image analysis of T59 binding with PPyCl was carried out under the same conditions
as for titer count analysis (106 PFU per sample). Phage binding was qualitatively analysed using an antiM13 bacteriophagebiotin conjugate antibody (Sigma, St. Louis, Missouri), followed by labelling with
streptavidin-conjugated fluorescein isothiocyanate (FITC) (1:500 in PBS, Sigma). Samples were imaged
immediately using a confocal laser scanning microscope (Leica TCS 4D, 488-nm FITC excitation).
T59 peptide corresponding to the phage was synthesized (ICBM, UT-Austin) with a Gly-GlyGly-Ser (GGGS) linker for stability at the C-terminus, making the peptide sequence THRTSTLDYFVIGGGS-K-biotin. 15 M peptide in PBS was incubated with PPyCl, PS, and PPyPSS. The samples were
washed and labelled with streptavidinFITC for confocal analysis.
For quantitative binding assessment of T59 peptide and its variants (Supplementary
Information, Table SIII) to PPyCl, peptides without the presence of biotin and linker
(THRTSTLDYFVI) were synthesized (Biopolymers Lab, MIT) and used for analysis using
fluorescamine (4-phenylspirol[furan-2(3H),1-phthalan]-3,3dione) reagent (Molecular Probes,
Eugene, Oregon)32. PPyCl samples (size ~0.5 cm2) were incubated with 15 M T59 peptide (or
variant) in 10 mM PBS at room temperature for 1 h. For binding studies of T59 peptide in varying
pH conditions, PPyCl and T59 peptides were incubated in different pH (010) 10 mM PBS solutions.
The unbound peptide solution and a mass balance approach (input = bound + unbound) were used
to evaluate the actual bound peptide to PPyCl.
Relative binding strengths of streptavidinbiotin and T59-PPyCl were evaluated using AFM
(Asylum Research 3D MFP, Santa Barbara, California). Force measurements were taken at constant
loading rates (vertical piezo velocity of 500 nm s1) where the spring constant of the tip was calibrated
in PBS by the thermal fluctuation method46 (12 2 pN nm1). The most probable unbinding force
was determined by fitting a gaussian fit to the histogram generated by the force distribution (n = 300,
10 spots per sample and 30 curves per spot)37,39 (data not shown). For AFM tip preparation, Si3N4
cantilevers (Veeco, Santa Barbara, California) were functionalized using streptavidinbiotin and
biotinylated peptides (GRGDSbiotin, T59-biotin, T59GRGDSbiotin)47.
T59 peptide was synthesized (Biopolymers Lab, MIT) with a cell-adhesion-promoting sequence
(GRGDS) at the C-terminus, THRTSTLDYFVI-GRGDS (T59-RGDS). PPyCl films were affixed with
Plexiglas chambers and sterilized using ultraviolet light. Peptides (15 M) were incubated to PPyCl for
1 h in PBS, followed by rinsing. PC12 cells (ATCC) in F12K medium were seeded onto each substrate
with 15% horse serum/2.5% foetal bovine serum (positive control) or serum-free medium at a density
of 104 cells cm2. Both live/dead stain and MTS assay were used to evaluate cell adhesion.
In vitro physiological stability of PPyCl-T59 was assessed in 10% serum-containing phenol redfree Dulbeccos Modified Eagles Medium (DMEM). PPyCl samples were incubated with 15 M T59,
followed by incubation in DMEM and 10 mM PBS for up to 21 days at 37 C. A fluorescamine assay was
used to evaluate binding.

Received 6 April 2004; accepted 15 April 2005; published 15 May 2005.

References
1. Zhang, S. & Altman, M. Peptide self-assembly in functional polymer science and engineering. React.
Funct. Polym. 41, 91102 (1999).
2. Valuev, I. L., Chupov, V. V. & Valuev, L. I. Chemical modification of polymers with physiologically
active species using water-soluble carbodiimides. Biomater. 19, 4143 (1998).
3. Xiao, S. J., Textor, M. & Spencer, N. Covalent attachment of cell-adhesive, (Arg-Gly-Asp)- containing
peptides to titanium surfaces. Langmuir 14, 55075516 (2000).
4. Shakesheff, K., Cannizzaro, S. & Langer, R. Creating biomimetic micro-environments with synthetic
polymer-peptide hybrid molecules. J. Biomater. Sci. Polym. E 9, 507518 (1998).
5. Tong, Y. W. & Shoichet, M. S. Peptide surface modification of poly(tetrafluoroethylene-cohexafluoropropylene) enhances its interaction with central nervous system neurons. J. Biomed.
Mater. Res. 42, 8595 (1998).
6. Hrkach, J., Ou, J., Lotan, N. & Langer, R. Poly (L-lactic acid-co-amino acid) graft copolymers.
J. Mater. Sci. Res. 78, 92102 (1999).
7. Cook, A. D., Pajvani, U. B., Hrkach, J. S., Cannizzaro, S. M. & Langer, R. Colorimetric analysis of
surface reactive amino groups on poly(lactic acid-co-lysine):poly(lactic acid) blends. Biomater. 18,
14171424 (1997).
8. Kubies, D., Rypacek, F., Kovarova, J. & Lednicky, F. Microdomain structure in polylactide-blockpoly(ethylene oxide) copolymer films. Biomater. 21, 529536 (2000).

501

nature materials | VOL 4 | JUNE 2005 | www.nature.com/naturematerials

2005 Nature Publishing Group

nmat1397-print.indd 501

2005 Nature Publishing Group

9/5/05 4:47:29 pm

ARTICLES
9. Mori, T. Cancer-specific ligands identified from screening of peptide-display libraries. Curr. Pharm.
Des. 19, 23352343 (2004).
10. Rowley M. J., OConnor, K. & Wijeyewickrema, L. Phage display for epitope determination: A
paradigm for identifying receptor-ligand interactions. Biotechnol. Annu. Rev. 10, 15188 (2004).
11. Whaley, S. R., English, D. S., Hu, E. L., Barbara, P. F. & Belcher, A. M. Selection of peptides with
semiconductor binding specificity for directed nanocrystal assembly. Nature 405, 665668 (2000).
12. Flynn, C. E. et al. Synthesis and organization of nanoscale II VI semiconductor materials using
evolved peptide specificity and viral capsid assembly. J. Mater. Chem. 13, 24142421 (2003).
13. Sano, K. & Shiba, K. A hexapeptide motif that electrostatically binds to the surface of titanium.
J. Am. Chem. Soc. 125, 1423414235 (2003).
14. Sarikaya, M., Tamerler, C., Jen, A. K., Schulten, K. & Baneyx F. Molecular biomimetics:
nanotechnology through biology. Nature Mater. 2, 577585 (2003).
15. Sarikaya M., Tamerler C., Schwartz D. T. & Baneyx F. Materials assembly and formation using
engineered polypeptides. Annu. Rev. Mater. Res. 34, 373408 (2004).
16. Adey, N. B., Mataragnon, A. H., Rider, J. E., Carter, J. M. & Kay, B. K. Characterization of phage that
bind plastic from phage-displayed random peptide libraries. Gene 156, 2731 (1995).
17. Berglund, J., Lindbladh, C., Nicholls, I. A. & Mosbach, K. Selection of phage display combinatorial library
peptides with affinity for a yohimbine imprinted methacrylate polymer. Anal. Commun. 35, 37 (1998).
18. Konturri, K., Pentti, P. & Sundholm, G. Polypyrrole as a model membrane for drug delivery.
J. Electroanal. Chem. 453, 231238 (1998).
19. Schmidt, C. E., Shastri, V. R., Vacanti, J. P. & Langer, R. Stimulation of neurite outgrowth using an
electrically conducting polymer. Proc. Natl Acad. Sci. USA 94, 89488953 (1997).
20. Valentini, R. F., Vargo, T. G., Gardella, J. A. Jr. & Aebischer, P. Electrically conductive polymeric
susbtrates enhance nerve fibre outgrowth in vitro. Biomater. 13, 183190 (1992).
21. Vidal, J. C., Garcia, E. & Castillo, J. R. In situ preparation of a cholesterol biosensor: entrapment
of cholesterol oxidase in an overoxidized polypyrrole film electrodeposited in a flow system:
Determination of total cholesterol in serum. Anal. Chim. Acta. 385, 213222 (1999).
22. Cui, X., Hetke, J. F., Wiler, J. A., Anderson, D. J. & Martin, D. C. Electrochemical deposition and
characterization of conducting polymer polypyrrole/PSS on multichannel neural probes. Sensors
Actuat. A 93, 818 (2001).
23. Wang, L., Li, X. & Yang, Y. Preparation, properties and applications of polypyrroles. React. Funct.
Poly. 47, 125139 (2001).
24. Sutar, D., Menon, R. & Subramanyam, S. V. Study of electrical conduction in polypyrrole by varying
doping level. Thin Solid Films. 417, 4042 (2002).
25. Silk, T., Hong, Q., Tamm, J. & Compton, R. G. AFM studies of polypyrrole film surface morphology
I. The influence of film thickness and dopant nature. Synth. Met. 93, 5964 (1998).
26. Collier, J. H., Camp, J. P., Hudson, T. W. & Schmidt, C. E. Synthesis and characterization of
polypyrrole-hyaluronic acid composite biomaterials for tissue engineering applications. J. Biomed.
Mater. Res. 50, 574584 (2000).
27. Diaz, A. F., Castillo, J. A., Logan, J. A. & Lee, W. Y. Electrochemistry of conductive polypyrrole films.
J. Electroanal. Chem. 129, 115132 (1981).
28. Tamm, J., Hallik, A., Alumaa, A. & Sammelselg, V. Electrochemical properties of polypyrrole/
sulphate films. Electrochim. Acta 42, 29292934 (1997).
29. Lee, S.-W., Mao, C., Flynn, C. E. & Belcher, A. M. Ordering of quantum dots using genetically
engineered viruses. Science 296, 892895 (2002).
30. Technical Bulletin E8100, E8110, E8120 14 (New England Biolabs, Cambridge, Massachusetts, 2004).
31. Blackburn, S. Amino Acid Determination: Methods and Techniques 2nd edn, 121125 (Marcel Dekker,
New York, 1978).

32. Undenfriend, S. et al. Fluorescamine: A reagent for assay of amino acids, peptides, proteins, and
primary amines in the picomole range. Science.172, 871872 (1972).
33. Barrett, G. C. & Elmore, D. T. Amino acids and peptides (Cambridge Univ. Press, New York,
1998).
34. Lee, G. U., Kidwell, D. A. & Colton, R. J. Sensing discrete streptavidin-biotin interactions with
atomic force micrscopy. Langmuir 10, 354357 (1994).
35. Dammer, U. et al. Specific antigen/antibody interactions measured by force spectroscopy
Biophys. J. 70, 24372441 (1996).
36. Nakamura, C. et al. Mechanical force analysis of peptide interactions using atomic force
microscopy. Biopolym. (Peptide Sci.) 76, 4854 (2004).
37. Merkel, R., Nassoy, P., Leung, A., Ritchie, K. & Evans, E. Energy landscapes of receptor-ligand
bonds explored with dynamic force spectroscopy. Nature 397, 5053 (1999).
38. Allison, D. P., Hinterdorfer. P. & Han, W. Biomolecular force measurements and the atomic
force microscope. Curr. Opin. Biotech. 13, 4751 (2002).
39. Evans, E. Energy landscapes of biomolecular adhesion and receptor anchoring at interfaces
explored with dynamic force spectroscopy. Faraday Discuss. 111, 116 (1998).
40. Massia, S. P. & Hubbel, J. A. Covalently attached GRGD on polymer surfaces promotes
biospecific adhesion of mammalian cells. Ann. N. Y. Acad. Sci. 589, 261270 (1988).
41. Pierschbacher, M. D. & Ruoslahti, E. Cell attachment activity of fibronectin can be duplicated
by small synthetic fragments of the molecule. Nature 309, 3033 (1984).
42. Hersel, U., Dahmen, C. & Kessler, H. RGD modified polymers: biomaterials for stimulated cell
adhesion and beyond. Biomater. 24, 43854415 (2003).
43. Barltrop, J. A., Owen, T. C., Cory, A. H. & Cory, J. G. 5-(3-carboxymethoxyphenyl)-2-(4,5dimethylthiazoly)-3-(4 sulfophenyl)tetrazolium, inner salt (MTS) and related analogs of 3-(4,5dimethylthiazolyl)-2,5-diphenyltetrazolium bromide (MTT) reducing to purple water-soluble
formazans as cell-viability indicators. Bioorg. Med. Chem. Lett. 1, 611614 (1991).
44. Labrou, N. E. Design and selection of ligands for affinity chromatography. J. Chromatogr. B
790, 6778 (2003).
45. Jozefonvicz, J. & Jozefowicz, M. Review: Interactions of biospecific functional polymers with
blood proteins and cells. J. Biomater. Sci. Polym. E 1, 147165 (1990).
46. Lvy, R., & Maaloum, M. Measuring the spring constant of atomic force microscope cantilevers:
thermal fluctuations and other methods. Nanotechnol. 13, 3337 (2002).
47. Moy, V. T., Florin, E.-L. & Gaub, H. E. Adhesive forces between ligand and receptor measured by
AFM. Colloids Surf. A 93, 343348 (1994).

Acknowledgements
This work was supported by the Gillson Longenbaugh Foundation (C. E. S.) and the Welch
Foundation (C. E. S.), the Packard Foundation (A. M. B.) and the National Science Foundation
(A. M. B.). The authors wish to thank the Center for Nano and Molecular Science and Technology
for use of the AFM, and Wolfgang Frey for helpful discussions on peptide interactions and force
spectroscopy. We would also like to thank John Mendenhall and Klaus Linse at the Institute of
Cellular and Molecular Biology for their help with confocal microscopy and peptide synthesis (and
other peptide related discussions), respectively.
Correspondence and requests for materials should be addressed to C.E.S.
Supplementary Information accompanies the paper on www.nature.com/naturematerials.

Competing financial interests

The authors declare that they have no competing financial interests.

502

nature materials | VOL 4 | JUNE 2005 | www.nature.com/naturematerials

2005 Nature Publishing Group

nmat1397-print.indd 502

2005 Nature Publishing Group

9/5/05 4:47:31 pm