Coiled coil peptides as universal linkers for the attachment of recombinant proteins to polymer therapeutics

Article abstract of DOI:10.1021/bm200897b

We have designed, synthesized, and characterized peptides containing four repeats of the sequences VAALEKE (peptide E) or VAALKEK (peptide K). While the peptides alone adopt in aqueous solutions a random coil conformation, their equimolar mixture forms heterodimeric coiled coils as confirmed by CD spectroscopy. 5-Azidopentanoic acid was connected to the N-terminus of peptide E via a short poly(ethylene glycol) spacer. The terminal azide group enabled conjugation of the peptide with a synthetic drug carrier based on the N-(2-hydroxypropyl)methacrylamide copolymer containing propargyl groups using "click" chemistry. When incorporated into the polymer drug carrier, peptide E formed a stable noncovalent complex with peptide K belonging to a recombinant single-chain fragment (scFv) of the M75 antibody. The complex thereby mediates a noncovalent linkage between the polymer drug carrier and the protein. The recombinant scFv antibody fragment was selected as a targeting ligand against carbonic anhydrase IX - a marker overexpressed by tumor cells of various human carcinomas. The antigen binding affinity of the polymer-scFv complex was confirmed by ELISA. This approach offers a well-defined, specific, and nondestructive universal method for the preparation of protein (antibody)-targeted polymer drug and gene carriers designed for cell-specific delivery.

Full text of DOI:10.1021/bm200897b

Article
pubs.acs.org/Biomac
Coiled Coil Peptides as Universal Linkers for the Attachment of
Recombinant Proteins to Polymer Therapeutics
,
†
†
§
†
†
́ ́ ̌
Michal Pechar,* Robert Pola, Richard Laga, Karel Ulbrich, Lucie Bednarova, Petr Malon
,‡
‡
§
§
∥
́ ́ ́ ̌ ̌
Irena Sieglova, Vlastimil Kral, Milan Fabry, and Ondrej Vanek
†Institute of Macromolecular Chemistry, Academy of Sciences of the Czech Republic, Heyrovskeho nam. 2, 162 06, Prague 6, Czech
Republic
‡
Institute of Organic Chemistry and Biochemistry, Academy of Sciences of the Czech Republic, Flemingovo nam. 2, 166 10 Prague 6,
Czech Republic
§
Institute of Molecular Genetics, Academy of Sciences of the Czech Republic, Flemingovo nam. 2, 166 10 Prague 6, Czech Republic
∥Department of Biochemistry, Faculty of Science, Charles University in Prague, Hlavova 8, 128 40 Prague, Czech Republic
*
S Supporting Information
ABSTRACT: We have designed, synthesized, and charac-
terized peptides containing four repeats of the sequences
VAALEKE (peptide E) or VAALKEK (peptide K). While the
peptides alone adopt in aqueous solutions a random coil
conformation, their equimolar mixture forms heterodimeric
coiled coils as confirmed by CD spectroscopy. 5-Azidopenta-
noic acid was connected to the N-terminus of peptide E via a
short poly(ethylene glycol) spacer. The terminal azide group
enabled conjugation of the peptide with a synthetic drug
carrier based on the N-(2-hydroxypropyl)methacrylamide copolymer containing propargyl groups using “click” chemistry. When
incorporated into the polymer drug carrier, peptide E formed a stable noncovalent complex with peptide K belonging to a
recombinant single-chain fragment (scFv) of the M75 antibody. The complex thereby mediates a noncovalent linkage between
the polymer drug carrier and the protein. The recombinant scFv antibody fragment was selected as a targeting ligand against
carbonic anhydrase IXa marker overexpressed by tumor cells of various human carcinomas. The antigen binding affinity of the
polymer−scFv complex was confirmed by ELISA. This approach offers a well-defined, specific, and nondestructive universal
method for the preparation of protein (antibody)-targeted polymer drug and gene carriers designed for cell-specific delivery.
INTRODUCTION
clinical applications more easily. The specific supramolecular
noncovalent self-assembly of the two macromolecules
represents a feasible solution for the general problem.
■
Hybrid copolymers consisting of both synthetic and natural
macromolecules have become attractive materials for various
biomedical applications. Conjugation of biologically active
proteins (e.g., antibodies, enzymes, and antibody fragments) to
synthetic hydrophilic polymer carriers often improves the
pharmacokinetics of the proteins, prolongs their blood
circulation, reduces unwanted immunogenicity, slows down
proteolytic degradation, and increases the accumulation of
macromolecular therapeutics in solid tumors by an enhanced
1
A heterodimeric coiled coil interaction has been recently
6−9
reported
as a linkage between biologically active molecules
and synthetic hydrophilic polymers. In this work, we suggest
using a coiled coil motif as a universal tool to bind recombinant
proteins to synthetic hydrophilic polymers. We report here on
the unique combination of recombinant technology and
polymer chemistry for the preparation of a new class of well-
defined targeted polymer therapeutics. It has been shown that
attachment of a synthetic polymer to the coiled coil peptide
does not disturb the coiled coil domain formation. On the
contrary, the thermal stability of the polymer-modified coiled
2
permeation and retention (EPR) effect.
Antibodies and their fragments have been described as very
efficient ligands for the specific targeting of both solid tumors
3
−5
and leukemia cells. However, covalent attachment of the two
macromolecules (protein and synthetic polymer) is often
nonselective (multiple functional groups of the protein are
modified); the resulting polymer−protein conjugate is a
complex mixture of macromolecules with high polydispersity
and compromised biological activity. The precisely defined
structure of the conjugate is highly desirable to achieve the
optimal biological activity of the product. Last but not least,
well-defined chemical entities obtain regulatory approval for
10,11
coil was even increased.
We have designed and synthesized the peptides (VAA-
LEKE) (peptide E) and (VAALKEK) (peptide K), forming
4
4
coiled coil heterodimers in aqueous solution. Peptide K was
Received: June 30, 2011
Revised: August 19, 2011
Published: August 25, 2011
©
2011 American Chemical Society
3645
dx.doi.org/10.1021/bm200897b|Biomacromolecules 2011, 12, 3645−3655

Biomacromolecules
Article
incorporated into the structure of a recombinant single-chain
fragment (scFv) of the M75 antibody without changing the
structure of the binding part of scFv.
column, 100 × 4.6 mm (Merck, Germany), using precolumn
derivatization with phthalaldehyde and 3-sulfanylpropanoic acid
(
excitation at 229 nm, emission at 450 nm) and gradient elution
with 10−100% of solvent B for 35 min at flow rate of 1.0 mL/min
solvent A: 0.05 M sodium acetate buffer, pH 6.5; solvent B: 300 mL
12−14
The antibody is known
to specifically bind to carbonic
(
anhydrase IX (CA IX), which is a transmembrane protein
overexpressed in a wide variety of tumor cell types. Peptide E
was covalently bound to a synthetic hydrophilic copolymer
based on N-(2-hydroxypropyl)methacrylamide (HPMA) using
of 0.17 M sodium acetate and 700 mL of methanol). MALDI TOF
spectroscopy was carried out using a Bruker Biflex III mass
spectrometer.
Synthesis of Peptides and Their Derivatives (1−6). The
loading of the 2-chlorotrityl chloride resin with Fmoc-amino acid in
DCM in the presence of DIPEA (4 equiv) was assessed by the
spectrophotometric determination of the Fmoc group (0.37 mmol/g)
released with 25% piperidine in DMF. The linear fully protected
heptapeptide was assembled using an AVSP-2 multiple automatic
peptide synthesizer (Development Workshops of the Institute of
Organic Chemistry and Biochemistry, Academy of Sciences of the
Czech Republic), starting from the C-terminus using standard Fmoc
procedures, by consecutive addition of the N-Fmoc-protected amino
acid (2.5 equiv), PyBOP (2.5 equiv), HOBt (2.5 equiv), and DIPEA
(5.0 equiv) in DMF. The Fmoc groups were removed using
piperidine−DMF (1:4). The Fmoc-(Fmoc-Hmb)Ala-OH derivative
was used for the incorporation of the Ala residue following Leu in
peptides 3 and 4. The following Fmoc-Ala-OH (10 equiv) was coupled
“
click” chemistry. A stable noncovalent complex was formed
upon mixing the recombinant protein with the polymer under
physiological conditions due to the formation of intermolecular
coiled coil heterodimers. We have investigated receptor-specific
cell binding (to CA IX) of the polymer−protein complex using
ELISAs.
The combination of synthetic polymer and peptide chemistry
with recombinant DNA technology described in this paper
offers a new approach for the preparation of polymer−protein
conjugates for cell-specific therapy. The specific noncovalent
and nondestructive attachment of a recombinant protein (e.g.,
targeting antibody fragment) to a synthetic hydrophilic polymer
via the coiled coil interaction is the major advantage of the
presented method. The polymer−protein conjugates prepared
by the self-assembly method can be used as targeted polymer
antitumor therapeutics or diagnostics.
as its symmetric anhydride preformed with DIC (10 equiv) in 50%
Me,Me
DMF−DCM. Fmoc-Glu(OtBu)-Ser(ψ
pro)-OH and Fmoc-Lys-
Me,Me
(Boc)-Ser(ψ
pro)-OH pseudoproline dipeptides were used for the
synthesis of peptides 5 and 6, respectively. Cleavage of the protected
peptides from the resin was performed with a 30% solution of HFIP in
DCM, 25 mL/g of resin, for 2 h. The resin was filtered off and rinsed
with DCM, the filtrate was concentrated under vacuum, and the oily
EXPERIMENTAL SECTION
Materials and Methods. 1-Aminopropan-2-ol, 2,2′-azobis-
isobutyronitrile) (AIBN), N,N-dimethylformamide (DMF), dimethyl
■
(
residue was precipitated with Et O. The precipitate was isolated by
2
filtration and dried in vacuum. The yields were typically 90−99%
based on the resin substitution with the first amino acid. Protected
peptides 1−6 were characterized by MALDI-TOF MS (Table 1).
sulfoxide (DMSO), ethane-1,2-dithiol (EDT), 1,1,1,3,3,3-hexafluor-
opropan-2-ol (HFIP), glycylglycine, ethyldiisopropylamine (DIPEA),
1
-hydroxybenzotriazole (HOBt), methacryloyl chloride, phenol,
piperidine, thioanisole, trifluoroacetic acid (TFA), triisopropylsilane
TIPS), and all other reagents and solvents were purchased from
(
Table 1. Basic Characteristics of Protected Heptapeptides
Sigma-Aldrich (Czech Republic). 2-Chlorotrityl chloride resin,
TentaGel Rink amide resin, (benzotriazol-1-yloxy)-
trispyrrolidinophosphonium hexafluorophosphate (PyBOP), 9-fluore-
nylmethoxycarbonyl (Fmoc), tert-butoxycarbonyl (Boc), tert-butyl
1
−6
theor mol
measdc
mol wt
a
b
compd
structure
wt
(
tBu) amino acid derivatives, and α-O-(2-carboxyethyl)-ω-O-(2-
1
Fmoc-Val-Ala-Ala-Leu-Glu(OtBu)-
Lys(Boc)-Glu(OtBu)-OH
1192.66
1235.71
1370.73
1215.64
1258.68
1393.47
Fmoc-aminoethyl)tetraethylene glycol (Fmoc-Peg ) were purchased
from Iris Biotech, GmbH, Germany. 2-Hydroxy-4-methoxybenzyl
(
from Merck. α-O-(2-Carboxyethyl)-ω-O-(2-Fmoc-aminoethyl) hepta-
kosaethylene glycol (Fmoc-Peg ) was purchased from Polypure AS,
4
2
3
Fmoc-Val-Ala-Ala-Leu-Lys(Boc)-
Glu(OtBu)-Lys(Boc)-OH
Me,Me
Hmb) and pseudoproline (ψ
pro) amino acid derivatives were
Fmoc-Val-Ala-(AcHmb)Ala-Leu-
Glu(OtBu)-Lys(Boc)-Glu(OtBu)-
27
d
OH
Norway. All amino acids were of L-configuration. Methacryloyl
chloride, 1-aminopropan-2-ol, and dichloromethane were distilled
immediately before use. All chemicals and solvents were of analytical
grade. Solvents were purified and dried using standard procedures.
Monitoring of the conjugation of the peptide to the reactive
copolymer was performed by HPLC using column Chromolith
Performance RP-18e, 100 × 4.6 mm (Merck, Germany), and a linear
gradient of water−acetonitrile, 0−100% acetonitrile in the presence of
4
5
6
Fmoc-Val-Ala-(AcHmb)Ala-Leu-
1413.73
1091.58
1177.66
1436.50
1114.21
1200.23
d
Lys(Boc)-Glu(OtBu)-Lys(Boc)-OH
Fmoc-Val-Ala-Ala-Leu-Glu(OtBu)-
Me,Me
Ser(ψ
pro)-Glu(OtBu)-OH
Fmoc-Val-Ala-Ala-Leu-Lys(Boc)-
Me,Me
Ser(ψ
pro)-Lys(Boc)-OH
a
b
Sample numbering used in the text. Calculated monoisotopic
molecular weight of the peptide acid. M + Na relative molecular
weight values from MALDI-TOF analysis. AcHmb = 2-acetoxy-4-
methoxybenzyl.
c
0
.1% TFA with a UV−vis diode array detector (Shimadzu, Japan).
d
Determination of the molecular weights and polydispersities of the
copolymers was carried out by size exclusion chromatography (SEC)
on a HPLC system (Shimadzu) equipped with refractive index, UV,
and multiangle light scattering DAWN 8 EOS (Wyatt Technology
Corp., Santa Barbara, CA) detectors using a Superose 6 PC 3.2/30 or
Superose 12 PC 3.2/30 column (Pharmacia) and 0.05 M phosphate
buffer with 0.15 M NaCl, pH 6.5, at a flow rate of 0.08 mL/min. The
calculation of molecular weights from the light-scattering detector was
based on the known injected mass assuming 100% mass recovery. The
content of thiazolidine-2-thione (TT) groups was determined
spectrophotometrically on a Helios Alpha UV/vis spectrophotometer
Reverse phase HPLC of the crude product showed a single peak
Chromolith C18 column, gradient 50−100% acetonitrile in water for
0 min, 0.1% TFA, 2 mL/min, UV detection at 220 nm), indicating
(
1
that no further purification was necessary.
Fragment Condensation (Peptides 7−18). TentaGel Rink
amide resin (0.5 g, 0.1 mmol of Fmoc groups) was deprotected
with 25% PIP in DMF and washed with DMF and DMSO. Fully
protected heptapeptides 1−6 (0.05 mmol) were attached to the amino
group of the resin in DMSO using DIC (0.1 mmol) as a coupling
agent. The condensation was carried out for 6 h at 25 °C. The resin
was incubated with acetic anhydride (10 equiv) and DIPEA (20 equiv)
in DMF for 30 min to end-cap possible unreacted amino groups. The
(
Thermospectronic, UK) using the absorption coefficient for TT in
methanol, ε305 = 10 280 L mol− cm . The amino acid analysis of the
hydrolyzed samples (6 M HCl, 115 °C, 18 h in a sealed ampule) was
performed on a Chromolith Performance RP-18e reversed-phase
1
−1
3
646
dx.doi.org/10.1021/bm200897b|Biomacromolecules 2011, 12, 3645−3655

Biomacromolecules
Article
peptide resin was washed with DMF, deprotected with 25% PIP and
added 15 min later. After 15 min, the polymer was precipitated with
diethyl ether and isolated by centrifugation. The crude product was
dried, dissolved in water, purified on Sephadex G-25 (PD 10 column),
and lyophilized to yield 44 mg of white polymer 20. The biotin
2
% DBU in DMF, and washed again with DMF and DMSO. The
condensation (with 0.15 mmol of the protected heptapeptide),
acetylation, deprotection, and washing steps were repeated until the
desired number of heptad repeats (3 or 4) was assembled. Then the
peptide resin was condensed with Fmoc-Peg4 (2.5 equiv) using
PyBOP (2.5 equiv), HOBt (2.5 equiv), and DIPEA (5.0 equiv) in
DMF. After removal of Fmoc, the N-terminus was modified with
Fmoc-Cys(Trt)-OH (the Fmoc group was removed, and the N-
terminus was acetylated to obtain peptides 7−10) or 5-azidopentanoic
acid (for peptides 11−16) under the same conditions as described
above. The peptides were cleaved from the resin with a mixture of
1
8
content was 1.8% w/w, as determined by HABA assay. The
molecular weight parameters of polymer 20 are M = 28 000 and M /
w
w
M = 1.6.
n
Attachment of Coiled Coil Peptides to the Polymer
Precursor via “Click” Reaction: Example of Preparation (21,
2
2). Copolymer 20 (5 mg, 3 μmol of propargyl groups), peptide azide
1
3 (4 mg, 1.2 μmol), and sodium ascorbate (1.6 mg, 8 μmol) were
dissolved in 0.2 mL of water, and the solution was thoroughly bubbled
with Ar to remove oxygen. A 0.8 M oxygen-free solution of copper(II)
sulfate (4 μL, 3.2 μmol) was added. The progress of the reaction was
monitored by HPLC; all peptide azide was bound to the polymer
within 30 min. The polymer−peptide conjugate was separated by SEC
on Sephadex G-25 column in water and lyophilized to yield 7 mg of
polymer−peptide conjugate 21. The content of the peptide in the
copolymer as determined by amino acid analysis was 40% w/w.
Polymer−peptide conjugate 22 was prepared analogously starting with
peptide azide 14. The molecular weight parameters of polymer 21 are
M = 56 000, M /M = 1.7 and of polymer 22 are M = 55 000, M /
TFA−EDT−thioanisole−phenol−H O−TIPS (68.5:10:10:5.5:5:1)
2
(
1
peptides 7−10) or TFA−H O−TIPS (95:2.5:2.5) (peptides 11−
2
6). For peptides 17 and 18, the N-terminal valine residue was
modified with Fmoc-Peg27 in two consecutive condensation and
deprotection steps and acetylated. The crude peptides were purified on
a semipreparative Chromolith C18 column using a gradient elution of
acetonitrile−water with 0.1% TFA. For the basic characteristics of
peptides 7−18, see Table 2. Examples of HPLC chromatograms and
MALDI TOF MS spectra are shown in Figure S1.
w
w
n
w
w
M = 1.7.
n
Table 2. Basic Characteristics of Coiled Coil Peptides 7−18
Preparation of the Polymer−scFv Complex via Coiled Coil
Interactions. Polymer 21 was dissolved in PBS buffer (0.7 mg/mL)
and mixed with a solution of the fusion protein−scFv-K (0.7 mg/mL)
in volume ratios of 1:1, 1:4, and 1:8 (corresponding to molar ratios
theor mol
measd mol
a
b
c
d
compd
structure
wt
wt
7
8
9
Ac-Cys-Peg -(VAALEKE)
2588.40
2585.56
3370.82
3367.03
2610.45
2589.54
2586.98
3371.69
3368.79
2611.61
4
3
8
:1, 2:1, and 1:1 between the peptide E and the peptide K). The
Ac-Cys-Peg -(VAALKEK)
4
3
resulting complex was characterized by SEC, gel electrophoresis, and
analytical ultracentrifugation.
Ac-Cys-Peg -(VAALEKE)
4
4
1
1
0
1
Ac-Cys-Peg -(VAALKEK)
4
4
CD Spectroscopy. CD measurements were performed using
Jasco-815 dichrograph. The spectra were measured in 1 mm quartz cell
from 190 to 260 nm with a scanning speed of 20 nm/min, a response
time of 4 s, and two computer-averaged accumulations. The peptide
concentration was kept constant at 0.2 mg/mL in sodium phosphate
buffer. Temperature-dependent measurements were performed using a
PTC-423S/L Peltier type temperature control system in a 1 cm quartz
cell allowing for stirring of the measured solution (peptide
concentration 0.02 mg/mL). The measurement was performed at
222 nm to indicate the α-helical content (temperature interval 2−95
5-azidopentanoyl-Peg4-
(
VAALEKE)3
5-azidopentanoyl-Peg4-
VAALKEK)3
5-azidopentanoyl-Peg4-
VAALEKE)4
5-azidopentanoyl-Peg4-
VAALKEK)4
5-azidopentanoyl-Peg4-
VAALESE)4
5-azidopentanoyl-Peg4-
VAALKSK)4
Ac-Peg -Peg -(VAALEKE)
1
1
1
1
1
2
3
4
5
6
2607.61
3350.85
3347.06
3186.60
3179.02
2608.69
3351.75
3347.72
3185.81
3179.95
(
(
(
(
°
C, temperature gradient of 25 °C/h, response time of 32 s). The
(
spectra were expressed as mean molar ellipticity at 222 nm per residue
1
1
7
8
5628.21
5624.42
5626.65
5623.13
and were converted into α-helical fraction (f ) using a two-state
H
2
7
27
4
19
model.
Ac-Peg -Peg -(VAALKEK)
4
27
27
a
b
Analytical Ultracentrifugation. Sedimentation analysis was
performed using a ProteomeLabXL-I analytical ultracentrifuge
equipped with an An50Ti rotor (BeckmanCoulter). Compounds 17
and 18 were dissolved to a concentration of 0.4 mg/mL in 50 mM
sodium phosphate (pH 7.4) and 150 mM NaCl buffer, which was also
used as a reference, and the complex was prepared by mixing equal
volumes of the individual solutions. A sedimentation velocity
experiment was carried out at 48 000 rpm and 20 °C; absorbance
scans were recorded at 230 nm in 20 min intervals with 30 μm spatial
resolution. The buffer density and peptide partial specific volumes
were estimated in SEDNTERP 1.09. The partial specific volume of Peg
was estimated as 0.83 mL/g at 20 °C, and the partial specific volumes
of compounds 17 and 18 were then calculated on the basis of the
molar ratio of their peptide and Peg components. The data were
Sample numbering used in the text. Single letter amino acid
abbreviations of the peptide sequences are in parentheses. Ac = acetyl.
c
d
Calculated monoisotopic molecular weight of the peptide acid. M +
H relative molecular weight values from MALDI-TOF analysis.
Synthesis of Monomers. N-(2-Hydroxypropyl)methacrylamide
(
HPMA) was prepared by the reaction of methacryloyl chloride with
15
1
-aminopropan-2-ol in DCM. N-Methacryloylglycylglycine (Ma-GG-
OH) was prepared by the Schotten−Baumann acylation of
glycylglycine with methacryloyl chloride in aqueous alkaline
medium. 3-(N-Methacryloylglycylglycyl)thiazolidine-2-thione (Ma-
GG-TT) was prepared by the reaction of Ma-GG-OH with 4,5-
dihydrothiazole-2-thiol in DMF in the presence of N,N′-dicyclohex-
16
17
20
ylcarbodiimide.
analyzed with SEDFIT 12.1. The scFv-K protein and polymer 21
were dissolved to a concentration of 0.7 mg/mL in 20 mM MES (pH
Reactive Copolymer with TT Groups (19). The copolymer
poly(HPMA-co-Ma-GG-TT) was prepared by radical solution
copolymerization of HPMA (90 mol %) and Ma-GG-TT (10 mol
6
.7) and 300 mM NaCl buffer, which was also used as a reference. For
sedimentation analysis, equal volumes of protein and polymer
solutions or protein and buffer solutions were mixed. The
sedimentation velocity experiment was carried out at 45 000 rpm
and 20 °C; absorbance scans were recorded at 280 or 240 nm in 5 min
intervals with a spatial resolution of 30 μm. The sedimentation
equilibrium experiment was performed at 4−6−8−10−12−14 000
rpm at 4 °C. Absorbance data were collected at 280 nm by averaging
20 scans with a spatial resolution of 10 μm after 30 h (first scan) or 18
h (consecutive scans) of achieving equilibrium and were globally
%
) performed in DMSO at 50 °C for 6 h. The concentration of
monomers in the copolymerization mixture was 15% w/w, and that of
AIBN initiator was 2% w/w. The molecular weight parameters of
polymer 19 are M = 24 000 and M /M = 1.5.
17
w
w
n
Copolymer with Propargyl Groups and Biotin (20). N-(2-
Aminoethyl)biotinamide hydrobromide (1.1 mg, 3.0 μmol), polymer
1
dissolved in DMF (0.8 mL). Propargylamine (26 μL, 0.39 mmol) was
9 (50 mg, 39 μmol of TT groups), and DIPEA (1 μL, 5.8 μmol) were
3
647
dx.doi.org/10.1021/bm200897b|Biomacromolecules 2011, 12, 3645−3655

Biomacromolecules
Article
analyzed with SEDPHAT 8.2 using a noninteracting discrete species
21
model. The buffer density, scFv-K partial specific volume, and
dimensions of the sedimenting species were estimated in SEDNTERP.
The partial specific volume of polymer 21 was estimated as 0.80 mL/g
at 20 °C, and the average of the two values was used for the protein−
polymer complex. The sedimentation velocity experiment with the
scFv-K protein and the (VAALEKE) peptide was conducted in the
4
same way as for the free scFv-K protein, with the exception of the
addition of 1:5 molar excess of the peptide. Some continuous size
distributions were recalculated in SEDPHAT for better accuracy; these
are given directly in s20,w values corrected for temperature and solvent
contributions and are thus shifted to higher values compared to the
observed s* calculated in SEDFIT.
Construction of the Expression Vector for the Protein with
Fusion Sequence. Cloning of the scFv fragment derived from the
12
monoclonal antibody M75, scFv M75, has already been described.
To introduce peptide K, i.e., (VAALKEK) , at the C-terminus of the
scFv fragment, the 90 bp oligonucleotide duplex was synthesized as
four complementary oligonucleotides (K1 = 5′-gtactgtggcagcgctgaaa-
4
Figure 1. Helical wheel diagram depicting the peptide sequences
(VAALEKE) and (VAALKEK) designed to form an antiparallel
3 3
gagaaggttgcggccttgaaa; K2
= 5′gagaaggtggcggcactgaaa-
coiled coil heterodimer.
gaaaaggtcgccgctctgaaagagaagg; K3 = 5′tgccgccaccttctctttcaaggccg-
caaccttctctttcagcgctgccaca; K4 = 5′gtacccttctctttcagagcggcgaccttttcttt-
cag), where the 5′ ends of K2 and K3 are phosphorylated (with T4
polynucleotide kinase) and the 5′ ends of K1 and K4 contain four-base
overhangs allowing cloning into the Acc65I site contained in the scFv
M75 sequence between the myc-tag and C-terminal his-tag.
Oligonucleotides K1 + K3 and K2 + K4 were annealed, the resulting
duplexes were ligated, and a 90 bp oligoduplex was gel purified and
cloned into the Acc65I-opened vector containing the sequence of scFv
M75. Thus, the final construct codes for scFv M75 in the format VH-
determines the association number of the helices in the
superhelix. Placement of valine and leucine at the a and d
positions, respectively, promotes the formation of dimeric
coiled coils. The ionic interactions between residues in the e
and g positions of the two helices substantially contribute to the
stability of the coiled coil and, depending on the charge of the
side chain, may control the formation of either homo- or
heterodimers. Therefore, glutamic acid was placed at positions e
and g of one strand, and lysine occupied the e′ and g′ positions
of the other strand. The electrostatic attraction between the
oppositely charged residues drives the formation of the
heterodimeric coiled coil. Eventual homodimerization is
thermodynamically less favorable due to the electrostatic
repulsions between the like charges of the same sequence.
Positions b and c were filled with alanine, which is known for its
high α-helical propensity. Lysine and glutamic acid were placed
in the remaining f and f′ positions to balance the total charge of
the peptides and increase their solubility. We have also tried
serine in positions f and f′. Serine is hydrophilic and neutral in
terms of its helical propensity. However, the main reason for
the use of serine was the possibility of using so-called
pseudoproline dipeptides during the solid phase peptide
synthesis. These consist of a dipeptide in which a Ser (or
Thr) residue is protected as a proline-like oxazolidine.
Incorporation of these derivatives into the structure of a
protected peptide dramatically improves its solubility and, as a
result, improves the yield of the subsequent condensation
reactions.
(
gly ser) -V -myc-Kpeptide-His .
Expression and Purification of the Fusion Protein scFv-K.
4 4 L 5
For expression in E. coli BL21(DE3) cells, a modified pET-22(b)
vector was used in which the scFv coding sequence is preceded by the
PelB signal sequence, allowing for translocation of the product into the
periplasmic space, and followed by a His5 tag, allowing for product
isolation and purification by IMAC chromatography on Ni-CAM
(
Sigma). Final purification was achieved by ion exchange chromatog-
raphy on a MonoS column.
Receptor Binding ELISA. The binding activity of the polymer
1−scFv complex formed via coiled coil interactions was assayed in
2
the ELISA. The microtiter plates were coated overnight with the CA
IX antigen (100 ng per well in 100 μL of bicarbonate buffer pH 9.6),
blocked with 1% BSA in PBS, and treated with polymer 21−scFv
complexes prepared in molar ratios of 1:1, 1:4, and 1:8, respectively,
for three different protein concentrations: 10, 2, and 0.4 mg/L. ScFv-K
alone and unconjugated polymer 21 were used as the positive and
negative controls, respectively. The bound polymer and polymer−scFv
complex were detected with the streptavidin peroxidase (Px)
conjugate, whereas for the positive control, scFv-K, which lacks
biotin, the anti-pentaHis monoclonal antibody conjugated with
peroxidase was used. The color developed in the reaction catalyzed
by peroxidase was measured at 492 nm.
Although Figure 1 shows the two peptides in an antiparallel
orientation, we are aware of the possibility to obtain eventually
a parallel arrangement. We believe that in the final complex
formed between polymer 21 and the recombinant protein scFv-
K the antiparallel orientation would be enforced by spatial
requirements.
RESULTS AND DISCUSSION
■
Design and Synthesis of the Peptides. The repeating
heptapeptide sequences were chosen according to the general
2
2−24
rules described
for the formation of coiled coil
heterodimers. Briefly, the coiled coil motif contains two to six
α-helices forming a left-handed superhelical bundle. The
essential feature of the peptides forming the coiled coil motif
is a repeated heptapeptide sequence denoted a-b-c-d-e-f-g. A
helical wheel diagram (Figure 1) was used to design the peptide
sequences forming the intermolecular heterodimers used in this
work. The positions a and d are occupied with hydrophobic
residues, whereas the e and g positions are usually occupied
with charged or hydrophilic residues. The choice of the
hydrophobic amino acid residues at positions a and d
Protected heptapeptides 1−6 were prepared by standard
solid phase peptide synthesis using the Fmoc/t-Bu strategy on a
2-chlorotrityl chloride resin. Solid phase fragment condensation
of protected heptapeptides 1 and 2 (followed by N-terminal
modification) afforded peptides 7−14 with three or four heptad
repeats in relatively low yields (between 10 and 15% after
HPLC purification).
2-Hydroxy-4-methoxybenzyl (Hmb) derivatives of amino
25
acids have been reported to substantially improve yields and
the solubility of peptide sequences that are difficult to obtain.
3
648
dx.doi.org/10.1021/bm200897b|Biomacromolecules 2011, 12, 3645−3655

Biomacromolecules
Article
Scheme 1
The use of Hmb-protected heptapeptides 3 and 4 led to a
slightly increased yield (15−20%) of fragment condensation;
however, the benefit was paid off by more laborious and
expensive preparation of derivatives 3 and 4 compared with
protected heptapeptides 1 and 2.
confirmed by SEC (Figure S2). Therefore, we turned our
attention to another chemoselective conjugation technique,
alkyne−azide cycloaddition.
Synthesis of the polymer−peptide conjugates 21 and 22 was
based on the copper-catalyzed “click” reaction of peptide azides
Much more significant improvement in the efficiency of the
synthesis was achieved using oxazolidine derivatives 5 and 6.
Thus, peptides 15 and 16, containing the VAALESE and
VAALKSK sequences, respectively, were prepared by solid
phase fragment condensation in higher yield and purity (30−
1
3 and 14 with polymer precursor 20 containing propargyl
groups (Scheme 1). The alkyne−azide cycloaddition was a very
efficient and chemoselective method of binding unprotected
peptides to polymer carriers.
The HPMA-based copolymers described above contain
multiple peptides attached to the polymer backbone via a
short tetra(ethylene glycol) spacer. The polymer precursors 19
and 20 have a polydispersity of about 1.5, which is typical for
polymers prepared by classical radical polymerization. After
attachment of the multiple peptide sequences to the side chain
of the polymers, the resulting conjugates become even more
heterogeneous in terms of the number of peptides per one
polymer chain. Although this should not be a major problem
regarding the biological activity of the final product, its exact
physicochemical characterization is much more difficult.
Hence, we also prepared better defined poly(ethylene glycol)
derivatives of the coiled coil peptides as an alternative to the
polymer−peptide conjugates based on HPMA copolymers.
4
0% after HPLC purification). Unfortunately, the stability of
the corresponding coiled coil heterodimers, determined by a
measurement of the temperature dependence of the ellipticity
at 222 nm, was lower compared with the peptides based on the
original VAALEKE and VAALKEK repeating motifs.
Polymer−Peptide Conjugates. Initially, we intended to
attach the coiled coil peptides to the polymers using addition of
the cysteine thiol group of peptides 7−10 to pendant
maleimidyl groups incorporated in the polymer. Unfortunately,
we have encountered several difficulties with this approach.
First, it was quite difficult to remove all traces of thiol
scavengers from the crude peptide detached from the resin after
the solid phase synthesis, even with preparative HPLC.
These impurities also react with the maleimidyl groups, thus
significantly lowering the yield of the polymer−peptide
conjugate. Second, the maleimidyl groups also react, albeit
slowly, with the primary amino groups of the lysine residues.
This results in unwanted branching of the final product as we
Attachment of two Peg molecules to the coiled coil peptides
27
with four heptad repeats resulted in well-defined monodisperse
block copolymers 17 and 18 with molecular weights of 5628
and 5624, respectively (MALDI-TOF MS). The association
3
649
dx.doi.org/10.1021/bm200897b|Biomacromolecules 2011, 12, 3645−3655

Biomacromolecules
Article
Figure 2. CD spectra of the peptides 7 and 8 and their equimolar mixture (a) (solid line, 8; dashed line, 7; solid thick line, 7 + 8) and thermal
melting profiles in an equimolar mixture (b).
Figure 3. CD spectra of peptides 9 and 10 and their equimolar mixture (a) (solid line, 10; dashed line, 9; solid thick line, 9 + 10) and thermal
melting profiles in an equimolar mixture (b).
behavior of the two copolymers was studied using analytical
ultracentrifugation and CD spectroscopy.
208 nm was very close to 1, which is typical for a coiled coil
arrangement (Figure 3a).
Circular Dichroism. Although two-dimensional NMR
Attempts to use cysteinyl peptides 7−10 for conjugation to
the polymers were unsuccessful; therefore, we switched to
azide-terminated peptides 11−16. Peptides 11−12 containing
only three heptad repeats exhibited typical CD spectra for a
random coil conformation. Surprisingly, no significant changes
in the CD spectra were observed even after mixing their
equimolar amounts (Figure S3).
Analogous peptides 7 and 8, containing N-acetylcysteine
instead of 5-azidopentanoic acid as the terminal residue, have
shown a typical α-helix pattern. A possible explanation of this
phenomenon is the dimerization of the cysteinyl peptides via
disulfide bridges. We have observed spontaneous formation of
the disulfide in PBS buffer due to air oxidation using HPLC
analysis. The peak corresponding to the disulfide gradually rose
over time; however, it completely disappeared after the addition
of a large excess of the reducing agent tris(2-carboxyethyl)-
phosphine (TCEP). Unfortunately, we were not able to obtain
reliable CD spectra in the presence of TCEP; at the relevant
concentration of TCEP, the total UV absorption of the solution
below 220 nm was too high. We speculate that formation of the
disulfide contributes to the higher thermodynamic stability of
the coiled coil heterodimer.
9
spectroscopy has been recently used for detailed character-
ization of heterodimeric coiled coils, CD spectroscopy is still
the most important analytical tool for study of similar systems.
It has been used also in this work for the determination of the
secondary structure of the peptides, their mixtures, and
polymer−peptide conjugates. As peptides 7−18 show CD
spectra typical for an unfolded structure, we can assume that
formation of the α-helix upon addition of the other
complementary peptide is most likely caused by formation of
a coiled coil heterodimer consisting of the two complementary
peptides.
First, peptides 7 and 8 containing three heptapeptide repeats
were prepared. Their N-termini were modified with tetra-
(
ethylene glycol) spacer and N-acetylcysteine, enabling attach-
ment to polymers with maleimidyl groups. Although CD
spectra of these peptides alone did not show significant content
of helical structure, their equimolar mixture exhibited minima at
2
08 and 222 nm, which are typical for α-helix. Nevertheless, the
intensity of these minima is lower than could be expected for
typical α-helical conformation (Figure 2a).
Similar results were obtained with analogous peptides 9 and
1
0 consisting of four repeated heptads. In this case, the
proportions of α-helical fractions in the peptides 9 and 10 were
2% and 13%, respectively. Their mixture exhibited a
significantly higher α-helical content than the preceding sample
fH = 43%). The ratio of the mean residual ellipticity at 222 and
To increase the stability of the coiled coil heterodimers that
were eventually formed, peptides 13−18 containing four
heptad repeats were investigated. While the individual peptides
showed typical random coil like CD, upon mixing with the
complementary peptide, the CD spectra displayed α-helix
2
(
3
650
dx.doi.org/10.1021/bm200897b|Biomacromolecules 2011, 12, 3645−3655

Biomacromolecules
Article
Figure 4. CD spectra of the peptides and their equimolar mixtures: (a) solid line, 14; dashed line, 13; solid thick line, 13 + 14; (b) solid line, 16;
dashed line, 15; solid thick line, 15 + 16; (c) solid line, 16; dashed line, 13; solid thick line, 13 + 16; (d) solid line, 14; dashed line, 15; solid thick
line, 15 + 14.
pattern. This has been observed for the following pairs of
peptides: 13 + 14, 15 + 16, 17 + 18, 13 + 16, and 14 + 15
manifested by an increase in the Θ value. The inflex point
2
22
of the curve corresponds to the melting temperature of the α-
helix.
(
Figure 4).
Peptides 17 and 18 contain a monodisperse poly(ethylene
The coiled coil heterodimer prepared from peptides 7 and 8
consisting of three heptapeptide repeats exhibited a typical
melting curve with a melting temperature of 60 °C (Figure 2b).
An analogous mixture of peptides 9 and 10 (containing four
heptads) showed much higher thermal stability with a melting
point of 90 °C (Figure 3b). This finding can be explained by
the cooperative behavior of the coiled coil interaction.
Heterodimers based on azide-terminated peptides 11−16
revealed less stable but distinct helices for the mixture of the
longer peptides 13 and 14 (melting temperature of 61 °C,
Figure 5); no significant portion of α-helix for the shorter
peptides 11 and 12 was found. The thermal helix stability of 13
and 14 is almost the same as that of the mixture of shorter
peptides 7 and 8; however, it is much lower than that of the
analogous cysteinyl peptides 9 and 10. We attribute this fact to
the stabilization of the latter peptides by disulfide bridges
spontaneously formed in aqueous solutions by air oxidation.
A mixture of peptides 15 and 16 containing serine in the
heptapeptide repeating unit melted at 51 °C (Figure 5),
indicating a somewhat lower stability of the coiled coil
heterodimer compared to the coiled coil prepared from
peptides 13 and 14. The coiled coil heterodimers made of
glycol) chain consisting of 54 oxyethylene units. Their
association behavior was investigated using both CD spectros-
copy and analytical ultracentrifugation. The CD spectrum of
the equimolar mixture of the two peptides (Figure 6a) had a
18
typical α-helix shape. Moreover, the calculated α-helical
fraction in the mixture was 100%, which was significantly higher
than that in the mixtures of the other pairs of coiled coil
peptides. This finding is in accordance with the previously
1
0,11
reported
helix-stabilizing effect of Peg.
After proper characterization of the association behavior of
the coiled coil peptides 13−18, we decided to verify our initial
idea to use the coiled coil interaction for the noncovalent
connection of two large macromolecules using the polymer−
peptide conjugates 21 and 22. Upon mixing equimolar amounts
of the two polymers in PBS buffer, we observed a significant
increase in the molecular weight compared to the original
polymers (from 45 000 to 385 000), as measured by SEC with a
LS detector. At the same time, changes in the CD spectra were
evident, suggesting a high content of α-helical structure.
Surprisingly, polymer−peptide conjugate 22 alone has shown
a substantial degree of helicity; however, this was increased
further after the addition of polymer−peptide conjugate 21
“
cross-mixed” peptides 13 + 16 and 14 + 15 had melting
(
Figure 7a).
Melting Curves. The thermal stability of the coiled coil
temperatures of 55 and 53 °C, respectively (Figure 5). As a
result, we decided to use only the more stable VAALEKE/
VAALKEK repeating motif in subsequent preparations of the
peptide−polymer conjugates.
heterodimers was verified by measuring the temperature
dependence of the ellipticity at 222 nm. Loss of the secondary
helical structure in the peptide mixture upon heating is
3
651
dx.doi.org/10.1021/bm200897b|Biomacromolecules 2011, 12, 3645−3655

Biomacromolecules
Article
their mixture, indicating quantitative complex formation.
However, while the sedimentation equilibrium data proved
that the mixture of compounds 17 and 18 behaves like a
heavier particle than its components, all equilibrium data
displayed shifts resembling aggregation and were therefore not
evaluated (data not shown).
Noncovalent Polymer−scFv Antibody Fragment Con-
jugate Formed by Coiled Coil Interaction. For the
ultimate proof of the feasibility of the concept of noncovalent
conjugation of a targeting antibody fragment to the polymer
carrier via coiled coil interaction, we chose a recombinant
single-chain fragment of the monoclonal antibody M75
(
directed against human carbonic anhydrase IX) engineered
to contain the (VAALKEK) peptide near its C-terminus and
4
capable of forming a coiled coil heterodimer with the partner
Figure 5. Thermal melting profiles of equimolar mixtures of peptides
peptide (VAALEKE) present in the structure of the HPMA
4
1
3 and 14 (solid line); 15 and 16 (dashed-dotted line); 13 and 16
copolymer.
(
dotted line); and 14 and 15 (dashed line).
The construction and expression of the scFv M75 fragment
carrying peptide (VAALKEK) at its C-terminus (scFv-K) are
described in detail in the Materials and Methods section.
Briefly, into the already existing scFv M75 DNA, an
It has been mentioned that peptides 17 and 18 containing
4
poly(ethylene glycol) (Peg) with 54 repeating ethylene glycol
units were prepared to investigate the effect of the
monodisperse polymer on the associative behavior of the
coiled coil peptides. We have also studied the influence of the
short Peg chain on the thermal stability of the coiled coil.
oligonucleotide duplex coding for (VAALKEK) was intro-
4
duced between the myc epitope and C-terminal his-tag, and the
final construct in the pET-22(b) vector was expressed in E. coli
BL21(DE3) cells. Following two chromatography steps, the
purified scFv-K protein (32 158 Da) was characterized using
SDS-PAGE, gel filtration, and analytical ultracentrifugation.
The recombinant product is highly pure and homogeneous,
as shown by SDS-PAGE (Figure 9a, lane 2); however, the SEC
10,11
Although it has been reported
that attachment of Peg to
the coiled coil peptide increases the melting temperature of the
α-helix, we did not observe any significant difference in the
thermal stability of the coiled coil heterodimers when
comparing the “Pegylated” peptides 17 and 18 (melting at 57
C, Figure 6b) with the “non-Pegylated” peptides 13 and 14
melting at 61 °C, Figure 5). Although the presence of the Peg
chain did not increase the melting temperature of the coiled
coil peptides, the content of the α-helical fraction of the
Pegylated” coiled coil heterodimer was significantly higher
compared to the “non-Pegylated” heterodimer.
(
Figure 9b) and analytical ultracentrifugation (Figure S4)
°
(
indicate the formation of a homodimer. The sedimentation
coefficient of the scFv-K protein was determined by a
sedimentation velocity experiment as s20,w = 3.57 ± 0.1 S
(
Figure S4a). This value is higher than smax, the theoretical
“
maximum sedimentation coefficient for a spherical particle of a
given scFv-K monomer mass, indicating scFv-K oligomeriza-
tion. The easiest explanation of these data would be an
elongated dimer with approximate dimensions of 15 × 3 nm for
a cylinder-shaped molecule. Indeed, the sedimentation
equilibrium experiment (Figure S4b), enabling direct calcu-
lation of the mass of sedimenting species, resulted in a z-
average molecular mass of 61 399 ± 453 Da for a non-
interacting discrete species model, indicating formation of the
scFv-K dimer. As the recombinant scFv without the
In contrast, a mixture of the HPMA-based polymer−peptide
conjugates 21 and 22 exhibited higher thermal stability than
peptides without a polymer. Unfortunately, it is very difficult to
determine the exact melting temperature of the sample, as the
melting curve does not show any significant inflex point. It is a
practically linear dependence with a relatively mild slope
compared to the steep melting curves of peptides 13−18. This
19
finding is in accordance with previously reported work
describing the stabilization effect of the HPMA copolymer on
the coiled coil formation. The effect is likely caused by
cooperative interactions of multiple coiled coil peptides grafted
to the polymer chain.
(
VAALKEK) fusion sequence behaves as a monomer in
4
solution (Figure 9b), it is possible to conclude that the scFv-K
dimerization is promoted by the presence of the (VAALKEK)4
peptide in the fusion protein, possibly through its unexpected
homotypic interaction. This interpretation is also in agreement
with the finding that polymer−peptide conjugate 22 alone
showed a substantial degree of helicity. The addition of an
Analytical Ultracentrifugation. Sedimentation analysis
of compounds 17 and 18 and their complex was hampered
by the fact that both molecules are relatively small and do not
contain a chromophore moiety. Therefore, the time for
sedimentation analysis had to be prolonged substantially, and
measuring at a short wavelength was necessary. Both facts
negatively affected the resolution of the analysis. Nevertheless,
the sedimentation velocity experiment proved complex
formation in a rather convincing way (Figure 8). The
sedimentation coefficients s20,w determined for compounds 17
and 18 were 0.57 and 0.52 ± 0.1 S, respectively, whereas their
mixture exhibited a single peak of 0.84 ± 0.08 S in the
sedimentation coefficient distribution. Sedimentation coeffi-
cient values correspond to a molecular weight of approximately
excess of free peptide (VAALEKE) leads to the disruption of
4
the protein homodimer and the formation of the heterodimer
peptide−protein, as shown in the AUC experiment in Figure
S5. Surprisingly, we have found larger particles in the mixture
with a sedimentation coefficient of 4.00 S that might
correspond to a heterotrimeric complex consisting of two
protein molecules and one or more peptides.
The addition of scFv-K to polymer 21 (containing peptide
E) resulted in a product with a molecular weight significantly
exceeding that of both components, as shown by analytical
ultracentrifugation analysis (Figure S6). The data from the
sedimentation equilibrium experiment (Figure S6b) could not
6
± 1 kDa for both compounds 17 and 18 or 11 ± 2 kDa for
3
652
dx.doi.org/10.1021/bm200897b|Biomacromolecules 2011, 12, 3645−3655

Biomacromolecules
Article
Figure 6. CD spectra of peptides 17 and 18 and their equimolar mixture (a) (solid line, 18; dashed line, 17; solid thick line, 17 + 18) and thermal
melting profiles in an equimolar mixture (b).
Figure 7. CD spectra of copolymers 21 and 22 and their equimolar mixture (solid line, 22; dashed line, 21; solid thick line, 21 + 22) (a) and thermal
melting profiles in an equimolar mixture (b).
heterogeneity of the analyzed sample. This observation is in
perfect agreement with the results of the sedimentation velocity
experiment (Figure S6a), showing the formation of at least
three distinct particles with sedimentation coefficients of 4, 6,
and 8 S, with a smaller fraction of even larger conjugates (8−16
S). As shown in Figure 10, it is unlikely that the 4 S complex
particles could correspond to unbound scFv-K protein because
the 0.5 S difference between the sedimentation coefficients of
free scFv-K protein and 4 S particles of the scFv-K−polymer 21
complex is too large to be ascribed to calculation error. As
polymer 21 seems to be completely associated in complex
formation (Figure 10, the absence of polymer 21 peak at 2 S), it
is more likely that the scFv-K dimer dissociated completely due
to coiled coil heterodimer formation with polymer 21. The
different particles observed in the complex mixture represent
different polymer 21−scFv-K stoichiometries, with the
dominant one being the 1:1 complex corresponding to the 4
S particle and higher stoichiometries (more scFv-K molecules
bound to one polymer 21 molecule) corresponding to larger
particles.
Receptor Binding of the Polymer−Protein Complex.
The ELISA experiment (Figure 11) has shown that the scFv
M75−polymer complex binds efficiently to its cognate antigen
CA IX. This proves that the coiled coil interaction between
peptide sequences comprising four repeats of VAALKEK on
the part of the antibody fragment and VAALEKE on the part of
polymer can serve as a universal noncovalent and non-
destructive attachment of the targeting protein to the polymeric
carrier.
Figure 8. Sedimentation analysis of compounds 17 and 18 and their
noncovalent complex. Individual compounds and their equimolar
mixture were analyzed in an analytical ultracentrifuge as described in
the Materials and Methods section. The plot shows the normalized
superposition of the resulting distributions of the sedimentation
coefficients of the species sedimenting in the analyzed samples as
calculated with the Sedfit program.
be fitted with either a single or two species model. However,
the addition of a third independent particle into the calculation
resulted in good fit with approximate z-average molecular
weights in the ranges of 60−90, 120−150, and 200−500 kDa.
These numbers do not correspond to the true masses of the
sedimenting species, but they reflect the complexity or the
3
653
dx.doi.org/10.1021/bm200897b|Biomacromolecules 2011, 12, 3645−3655

Biomacromolecules
Article
Figure 9. Characterization of the recombinant scFv-K protein by SDS-PAGE (a) and SEC (b). SDS-PAGE was performed on a 15% polyacrylamide
gel. Lane 1: standards of molecular weights in kDa (BioRad); lane 2: purified scFv-K. SEC was performed on a Superose 12 PC 3.2/30 column
(
Pharmacia) in 0.05 M phosphate buffer with 0.15 M NaCl, pH 6.5, at a flow rate of 0.08 mL/min. Blue line: scFv-K; red line: scFv without the
fusion peptide K.
The presence of the K-peptide at the C-terminal of the
engineered scFv-K does not interfere with its ability to bind to
the CA IX antigen, nor does it interfere with his-tag recognition
by the anti-pentaHis antibody in the ELISA (Figure 11, scFv-K
bars). Figure 11 gives only a qualitative result of receptor
binding. It is impossible to quantitatively compare the binding
efficiency of the scFv-K protein alone and the polymer−protein
complex due to different methods of evaluation. While the
polymer−protein complex was detected using a biotin label on
the polymer, the scFv-K protein was detected with an anti-
pentaHis antibody.
Nevertheless, the ELISA has clearly shown that the
polymer−protein complex binds to the antigen due to the
specific interaction between the scFv-K ligand and CA IX. The
polymer alone (without the targeting ligand) was used as a
negative control and has shown no binding.
Figure 10. Comparison of normalized size distributions of individual
complex components as well as the complex analyzed at two different
wavelengths calculated with the Sedphat program as detailed in the
Materials and Methods section.
The multivalency of the polymer 21 allows binding of up to
eight scFv-K molecules per one polymer chain. The formation
of the complexes containing more then one scFv-K was
confirmed by sedimentation analysis (Figure 10). Surprisingly,
we have found no significant differences between the levels of
binding of polymer−protein complexes containing different
amounts of scFv-K. We hypothesize that this observation can
be explained by steric hindrance when more than one scFv-K
ligand is attached to one polymer chain. If one scFv-K is already
bound to the receptor, the other ligands cannot effectively
interact with another CA IX molecule.
We plan to prepare more homogeneous complexes using
semitelechelic polymers synthesized by controlled radical
polymerization techniques (e.g., RAFT, ATRP). More detailed
biological evaluation of the polymer−protein complexes will be
a subject of the upcoming publications.
Figure 11. CA IX antigen binding by the protein−polymer complex
evaluated by ELISA at three different concentrations. From left to
right: scFv-K alone (positive control, detection by anti-pentaHis-Px);
scFv-K + polymer 21 (1:1, 4:1, 8:1 w/w), and polymer 21 (negative
control, detection by streptavidin-Px).
CONCLUSIONS
■
A new water-soluble polymer carrier system for potential
targeted drug delivery has been designed, prepared, and
3
654
dx.doi.org/10.1021/bm200897b|Biomacromolecules 2011, 12, 3645−3655

Biomacromolecules
Article
̌
́
́ ́
(15) Ulbrich, K.; Subr, V.; Strohalm, J.; Plocova, D.; Jelinkova, M.;
characterized. A biotinylated HPMA-based copolymer contain-
ing pendant peptide sequences (VAALEKE)4 has been
synthesized. Recombinant protein consisting of the scFv
fragment of the M75 antibody and the fusion peptide sequence
̌
́
Rihova,
16) Rejmanova,
78, 2159−2168.
́
B. J. Controlled Release 2000, 64, 63−79.
(
́
P.; Labsky, J.; Kopecek, J. Makromol. Chem. 1977,
́
̌
1
̌
(
(
(
(
(
17) Subr, V.; Ulbrich, K. React. Funct. Polym. 2006, 66, 1525−1538.
(
VAALKEK) was expressed and isolated from E. coli. The
4
18) Green, N. M. Biochem. J. 1965, 94, C23.
protein formed a stable complex with the copolymer due to
formation of coiled coil heterodimers between the comple-
mentary peptides. The polymer−protein complex exhibited
specific binding to the immobilized antigen in the ELISA. We
believe that the presented HPMA copolymer−coiled coil
system can be used as a universal tool for the attachment of
recombinant targeting proteins to synthetic polymers, thus
providing a receptor-specific drug carrier suitable for neoplastic
treatment or diagnosis.
19) Rohl, C. A.; Baldwin, R. L. Methods Enzymol. 1998, 295, 1−26.
20) Schuck, P. Biophys. J. 2000, 78, 1606−1619.
21) Schuck, P. Anal. Biochem. 2003, 320, 104−124.
(22) Mason, J. M.; Arndt, K. M. ChemBioChem 2004, 5, 170−176.
(23) Apostolovic, B.; Danial, M.; Klok, H.-A. Chem. Soc. Rev. 2010,
39, 3541−3575.
(24) Yu, Y. B. Adv. Drug Delivery Rev. 2002, 54, 1113−1129.
(25) Quibell, M.; Turnell, W. G.; Johnson, T. J. Chem. Soc., Perkin
Trans. 1 1995, 2019−2024.
26) Yang, J. Y.; Xu, C. Y.; Wang, C.; Kopece
006, 7, 1187−1195.
(
̌
k, J. Biomacromolecules
2
ASSOCIATED CONTENT
Supporting Information
■
*
S
Examples of SEC and HPLC chromatograms, MALDI TOF
MS spectra and CD spectra of peptides, sedimentation analysis
data with residual plots. This material is available free of charge
via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
E-mail: pechar@imc.cas.cz.
■
*
ACKNOWLEDGMENTS
■
The work was supported by the Grant Agency of the Czech
Republic, grant 203/08/0543, and by the Ministry of Education
of the Czech Republic, grant IM4635608802, and in part by
project AV0Z50520514 awarded by the Academy of Sciences of
the Czech Republic. We especially acknowledge Anna Van
for excellent technical assistance in peptide synthesis.
̌ ́
kova
REFERENCES
■
(
(
1) Kopece
̌
k, J. Eur. J. Pharm. Sci. 2003, 20, 1−16.
2) Maeda, H.; Matsumura, Y. Crit. Rev. Ther. Drug Carrier Syst.
1
989, 6, 193−210.
3) Etrych, T.; Strohalm, J.; Kovar
Ulbrich, K. J. Controlled Release 2009, 140, 18−26.
4) Kovar M.; Mrkvan, T.; Strohalm, J.; Etrych, T.; Ulbrich, K.;
̌
́
́ ̌ ̌ ́ ́
, L.; Kabesova, M.; Rihova, B.;
(
(
́ ̌
,
̌
̌
́
Stastny,
5) Hongrapipat, J.; Kopeck
J. Mol. Pharmaceutics 2008, 5, 696−709.
6) Wu, K. S.; Liu, J. H.; Johnson, R. N.; Yang, J. Y.; Kopece
Angew. Chem., Int. Ed. 2010, 49, 1451−1455.
7) Apostolovic, B.; Deacon, S. P. E.; Duncan, R.; Klok, H.-A.
Macromol. Rapid Commun. 2011, 32, 11−18.
8) Apostolovic, B.; Klok, H.-A. Biomacromolecules 2010, 11, 1891−
895.
9) Deacon, S. P. E.; Apostolovic, B.; Carbajo, R. J.; Schott, A. K.;
́
M.; Rihova,
́
B. J. Controlled Release 2003, 92, 315−330.
(
̌
ova, P.; Liu, J.; Prakongpan, S.; Kopece
́
̌
k,
(
̌
k, J.
(
(
1
(
Beck, K.; Vicent, M. J.; Pineda-Luceda, A.; Klok, H.-A.; Duncan, R.
Biomacromolecules 2011, 12, 19−27.
(
̌ ́ ̌
10) Pechar, M.; Kopeckova, P.; Joss, L.; Kopecek, J. Macromol.
Biosci. 2002, 2, 199−206.
11) Vandermeulen, G. W. M.; Tziatzios, C.; Klok, H.-A.
Macromolecules 2003, 36, 4107−4114.
12) Zavada, J.; Zavadova, Z.; Pastorek, J.; Biesova,
Velek, J. Br. J. Cancer 2000, 82, 1808−1813.
́
̌ ̌
13) Kral, V.; Mader, P.; Collard, R.; Fabry, M.; Hore , M.;
(
(
́
́
́
́
̌
Z.; Jezek, J.;
(
́
jsi
́ ́ ̌
ada, J.; Sedlacek, J.; Rulisek, L.; Brynda,
̌
́
k, M.; Zav
̌
Rezac
J. Proteins 2008, 71, 1275−1287.
14) Li, Y.; Wang, H.; Oosterwijk, E.; Selman, Y.; Mira, J. C.;
́
o
̌
va,
́
P.; Kozi
̌
se
(
Medrano, T.; Shiverick, K. T.; Frost, S. C. Biochem. Biophys. Res.
Commun. 2009, 386, 488−492.
3
655
dx.doi.org/10.1021/bm200897b|Biomacromolecules 2011, 12, 3645−3655