Squaric acid mediated synthesis and biological activity of a library of linear and hyperbranched poly(glycerol)-protein conjugates

Article abstract of DOI:10.1021/bm300103u

Polymer-protein conjugates generated from side chain functional synthetic polymers are attractive because they can be easily further modified with, for example, labeling groups or targeting ligands. The residue specific modification of proteins with side

Full text of DOI:10.1021/bm300103u

Article
pubs.acs.org/Biomac
Squaric Acid Mediated Synthesis and Biological Activity of a Library
of Linear and Hyperbranched Poly(Glycerol)−Protein Conjugates
Frederik Wurm,† Carsten Dingels,‡ Holger Frey,‡ and Harm-Anton Klok*,†
†
́
Ecole Polytechnique Fed
́
er
́
ale de Lausanne (EPFL), Institut des Mater
́ ́
iaux and Institut des Sciences et Ingenierie Chimiques,
Laboratoire des Polymer
̀
es, Bat
̂
iment MXD, Station 12, CH-1015 Lausanne, Switzerland
‡Department of Organic and Macromolecular Chemistry, Johannes Gutenberg-Universitat Mainz, Duesbergweg 10-14; 55099 Mainz,
̈
Germany
S
* Supporting Information
ABSTRACT: Polymer−protein conjugates generated from
side chain functional synthetic polymers are attractive because
they can be easily further modified with, for example, labeling
groups or targeting ligands. The residue specific modification of
proteins with side chain functional synthetic polymers using
the traditional coupling strategies may be compromised due to
the nonorthogonality of the side-chain and chain-end func-
tional groups of the synthetic polymer, which may lead to side
reactions. This study explores the feasibility of the squaric acid
diethyl ester mediated coupling as an amine selective, hydroxyl
tolerant, and hydrolysis insensitive route for the preparation of
side-chain functional, hydroxyl-containing, polymer−protein
conjugates. The hydroxyl side chain functional polymers selected
for this study are a library of amine end-functional, linear,
midfunctional, hyperbranched, and linear-block-hyperbranched polyglycerol (PG) copolymers. These synthetic polymers have
been used to prepare a diverse library of BSA and lysozyme polymer conjugates. In addition to exploring the scope and
limitations of the squaric acid diethylester-mediated coupling strategy, the use of the library of polyglycerol copolymers also
allows to systematically study the influence of molecular weight and architecture of the synthetic polymer on the biological
activity of the protein. Comparison of the activity of PG−lysozyme conjugates generated from relatively low molecular weight
PG copolymers did not reveal any obvious structure−activity relationships. Evaluation of the activity of conjugates composed of
PG copolymers with molecular weights of 10000 or 20000 g/mol, however, indicated significantly higher activities of conjugates
prepared from midfunctional synthetic polymers as compared to linear polymers of similar molecular weight.
with a single branching point bearing the protein-reactive group
in the center of a linear polymer chain leads to improved circula-
tion times as compared to conjugates based on linear PEG chains.7
Today, several FDA-approved PEGylated proteins are used for the
treatment of cancer, hepatitis C, anemia, and diabetes.8,9 A com-
mercial therapeutic for the treatment of hepatitis C (“PEGASYS,
Roche”) is available that relies on a midfunctional PEG coupled to
interferon α-2a.10
Although it is a highly valuable concept, protein PEGylation
also has several limitations: (i) linear PEG only possesses two
functional groups at the polymer chain ends, which restricts the
possibilities for further functionalization and the preparation of
multiprotein conjugates, and (ii) PEGylation often leads to a
loss of protein activity.11 The development of strategies to over-
come these limitations and that allow access to protein−polymer
INTRODUCTION
■
In recent years, with the discovery of novel pharmaceutically
active peptides and proteins, a novel field of human therapy is
rapidly evolving based on complex biopolymers.1 A drawback of
protein-based “biotherapeutics” is their fast degradation in the
human body by both the digestive system and the circulatory
system, that is, they are rapidly removed by proteolytic digestion
and renal excretion.2,3 In the last decades, the covalent attachment
of synthetic macromolecules has proven to be an effective strategy
to improve protein stability, reduce immunogenicity, extend
plasma half-life times, and increase solubility.4
Seminal work has been conducted in the 1970s by Davis and
Abuchowski, who found that covalent attachment of poly-
(ethylene glycol) (PEG) to bovine serum albumin (BSA) and
bovine liver catalase resulted in reduced immunogenicity and
increased blood circulation times.5,6 To date, among a variety of
other water-soluble and biocompatible polymers, PEG is still
the most frequently used for the modification of peptides and
proteins. In more recent work, it has been demonstrated that
the use of “branched” PEG-structures, that is, PEG derivatives
Received: January 19, 2012
Revised: February 19, 2012
Published: March 1, 2012
© 2012 American Chemical Society
1161
dx.doi.org/10.1021/bm300103u | Biomacromolecules 2012, 13, 1161−1171

Biomacromolecules
Article
conjugates with superior properties is a very topical and active
field of research.
group, which has been successfully used by Tao et al. for the
synthesis of PHPMA−protein conjugates,12 has a hydrolysis
half-life of several hours as compared to minutes for NHS
esters.24,25 The thiazolidine-2-thione group, however, has been
reported to undergo side reactions with thiol groups.25 Another
potentially interesting class of reagents to allow amine-selective
modification of proteins are squaric acid dialkyl esters, such as
squaric acid diethyl ester (SADE). While SADE has found
widespread use for glycosylation26−28 or other coupling reac-
tions between low molecular weight compounds,29,30 it has not
been explored for the preparation of synthetic polymer−protein
conjugates. Squaric acid dialkyl esters react selectively with
amine groups at pH 7 and room temperature without inter-
ference of transesterification or hydrolysis reactions.30 Due to
the lower reactivity of the reaction product as compared to the
starting material, only squaric acid ester monoamides are
formed, which can be isolated and characterized. Under more
basic reaction conditions (pH 9), the ester amide intermediates
can then be conjugated to another amine functional compound
in a second amine-selective and hydroxy-tolerant reaction,26
which makes the squaric acid mediated coupling a potentially
attractive approach to synthesize PG and other side chain
functional (co)polymer−protein conjugates.
In a recent study, we have explored the one-step activation of
amino poly(ethylene glycol)s (PEGs) with squaric acid diethyl
ester and investigated the reactivity and chemoselectivity of
these PEG derivatives toward the amine-selective modification
of α-amino acids and proteins.31 The present study is aimed at
further exploring and expanding the scope of this approach and
(i) explores the feasibility of the squaric acid mediated coupling
for the synthesis of side-chain functional synthetic polymer
(specifically, linear and hyperbranched polyglycerol)−protein
conjugates; (ii) uses the versatility of the glycidol ring-opening
polymerization to generate an architecturally diverse library of
PG−protein conjugates; and (iii) evaluates the biological activity
of these polymer−protein hybrids (Scheme 1).
One approach to generate “PEG-like” polymers with a large
number of functional groups is based on the controlled radical
polymerization of appropriate side-chain functional (meth)-
acrylates, such as 2-hydroxyethyl(meth)acrylate, poly(ethylene
glycol)(meth)acrylate, or 2-hydroxypropylmethacrylamide.12−14
These polymers can be conjugated to the peptide or protein of
interest via a “grafting from” or a “grafting to” approach to the
respective amino acid residues.14,15 Another interesting class of
PEG-like polymers, which, however, so far has been barely
explored for the synthesis of protein−polymer conjugates, are
polyglycerols (PGs).16 PGs are attractive because they are
biocompatible and nontoxic,17,18 as well as more protein-
resistant and thermally and oxidatively stable as compared to
PEG.19 PG can be obtained via ring-opening polymerization of
glycidol or glycidyl ethers in a controlled manner.18−22
Depending on the type of monomer used and the polymerization
conditions applied, both linear and (hyper)branched PGs can be
synthesized. This is attractive as it allows access to libraries of
chemically identical but structurally different PG analogues, which
may be valuable to investigate the influence of polymer
architecture on the properties of polymer−protein conjugates.
In many instances, PEGylation is aimed at modifying amine
groups of the peptide/protein of interest. Traditionally, site-
selective modification of amine groups is accomplished by
using, for example, N-hydroxysuccinimide (NHS), 4-nitro-
phenyl chloroformate, or similar end-functionalized PEG
derivatives.15,23 While these approaches work well for a large
variety of synthetic polymers, they cannot be applied to side-
chain (hydroxy) functional polymers such as PG, because the
classical active esters are susceptible to concurrent transesteri-
fication and hydrolysis. Lysine-selective protein modification
with PG (co)polymers requires protein-reactive groups that are
tolerant toward hydroxy groups and have extended hydrolysis
half-lives. One example of an amine-reactive, hydroxy-tolerant
activation group is the thiazolidine-2-thione group. This activating
Scheme 1. (Top) Squaric Acid Mediated Coupling of PG (Co)Polymers to Proteins; (Bottom) Schematic Representation of the
Architecturally Different Protein−Polymer Conjugates Prepared and Investigated in This Study
1162
dx.doi.org/10.1021/bm300103u | Biomacromolecules 2012, 13, 1161−1171

Biomacromolecules
Article
p-methoxybenzyl bromide (5.03 g, 25 mmol), potassium carbonate
(4.8 g, 35 mmol), and 80 mL DMF. Yield: (3.1 g, 78%). H NMR
EXPERIMENTAL SECTION
■
1
Materials. Ethylene oxide (Fluka 99%), glycidol (Acros 96%),
cesium hydroxide monohydrate (99.95% Aldrich), ethyl vinyl ether
(98% Aldrich), ethanolamine (99% Acros), serinol (2-amino-1,3-pro-
panediol, 98%, Acros), p-methoxybenzylalcohol (98% Acros), phos-
phorus tribromide (phosphorus tribromide, 1.0 M solution in dichloro-
methane, AcroSeal), potassium carbonate (99+% Acros), palladium on
activated charcoal (10%Pd, Aldrich), benzene (99% Acros), lysozyme
(90%), bovine serum albumin (98%), Micrococcus lysodeikticus
(lyophilized cells), and ninhydrin reagent were purchased from Aldrich
and used as received. Squaric acid diethyl ester was purchased from VWR
and used as received. α-Amino-ω-methoxy-poly(ethylene glycol)s were
purchased from IRIS Biotech and used as received. DMSO-d6 and CDCl3
were purchased from Deutero GmbH. Ethoxy ethyl glycidyl ether was
prepared as described by Fitton et al.,32 dried over CaH2, and freshly
distilled before use. Ethylene oxide and glycidol were distilled from CaH2
before use. p-Methoxybenzylbromide was synthesized as described
previously.33 All other reagents and solvents were purchased from Aldrich
and used as received, if not otherwise mentioned.
(300 MHz, DMSO): δ (ppm) 7.23−6.84 (8H, aromatic), 4.28 (br, 2H,
OH), 3.70 (s, 6H, OCH3), 3.62 (s, 4H, NCH2Ph), 3.54 (m, 2H,
CH2OH), 2.66 (t, J = 6 Hz, 1H, NCH). ESI-MS: 322.8 (MH+).
General Procedure for the Synthesis of Linear Poly(ethoxy ethyl
glycidyl ether) (Co)Polymers. The appropriate initiator (ca. 100 mg)
was placed in a Schlenk flask and dissolved in benzene (ca. 5 mL)
under an argon atmosphere. Cesium hydroxide monohydrate was
introduced to achieve a degree of deprotonation of 50%. The mixture
was stirred at 60 °C over a period of 45 min and then heated to 80−90 °C
in vacuo for 2 h to remove the formed water and benzene azeotropically. In
a separate setup, first dry THF, and the respective monomers (ethylene
oxide or ethoxy ethyl glycidyl ether (EEGE), predried over CaH2) were
distilled into a Schlenk flask equipped with a stir bar, a Teflon tap, and a
rubber septum. The flask was closed under vacuum and the cesium salt of
the initiator was added dissolved in dry DMSO to afford about a 10 wt %
solution of the monomer(s) in a 9:1 mixture of THF and DMSO. The
reaction mixture was directly heated to 70 °C in static vacuo over a period
of 20 h and quenched by the addition of about 20 mL of methanol and
0.5 g acidic ion-exchange resin. After that, the solution was filtered,
concentrated, and precipitated into diethyl ether/acetone (70:30) first and
subsequently in pure diethyl ether to yield the final polymer after drying in a
yield typically between 90% to quantitative. Note: PEEGE-homopolymers
1
Methods. H NMR spectra were recorded using either a Bruker
AC 300 (300 MHz spectra) or a Bruker AMX 400 instrument (400
MHz spectra). All spectra were referenced internally to the residual
proton signals of the deuterated solvent. SEC analysis in water was
carried out on a Viscotek TDA 300 instrument equipped with a
MetaChem degasser, a Visotek VE 1121 SEC solvent pump, and a VE
5200 SEC autosampler. A 9:1 mixture of phosphate buffer (0.1 M,
pH = 6.5)/methanol was used as a mobile phase. Samples were eluted
at 25 °C and at a flow rate of 0.5 mL·min−1 over Shodex OHpak 804
and 805 columns. Sample elution was monitored using a triple detection
setup and absolute molecular weights were determined via light scattering.
For SEC measurements in DMF (containing 0.25 g/L of lithium bromide
as an additive), an Agilent 1100 Series instrument equipped with a PSS
HEMA column (106/105/104 g/mol), a UV (275 nm), and a RI detec-
tor was used. Calibration was carried out using poly(ethylene oxide) stan-
dards provided by Polymer Standards Service. MALDI-ToF mass spec-
trometry was performed on a Shimadzu Axima CFR MALDI-ToF MS
mass spectrometer equipped with a nitrogen laser delivering 3 ns laser
pulses at 337 nm. α-Cyano-hydroxycinnamic acid (CHCA) was used as a
matrix and potassium triflate was added to facilitate ionization of polymer
samples. SDS-PAGE was carried out with 4−20% Tris-HCl gels (Biorad,
0.75 mm, 10 well). Reverse phase HPLC was performed on a Grace
Vydac C4-protein column using a modular setup from JASCO equipped
with a quaternary gradient pump PU-2089plus, autosampler AS-2055plus,
UV-detector UV-2075plus, and a column oven (25 °C) CO-2060plus.
Gradient elution was carried out at a flow rate of 0.5 mL/min with a
mobile phase A (99.9% H2O, 0.01% TFA) and a mobile phase B (99.9%
acetonitrile, 0.01% TFA). The gradient sequence (B) was 5−100% from
0−60 min, 100% from 60−80 min, 100−5% from 80−110 min. Sample
elution was monitored at a UV absorbance of 280 nm. Ninhydrin test: A
solution of the analyte was dropped onto a silica gel TLC plate, dipped
into the solution, and dried with a heatgun; a positive test was indicated
by a color change.
1
were filtered and precipitated twice into water. H NMR (for PEG-co-
PEEGE and PEG-b-PEEGE in DMSO-d6, 300 MHz): 7.22−6.82 (8H,
aromatic signals from the initiator), 4.64 (br s, acetal H), 3.69 (s, 6H, MeO-
C6H4-), 3.64−3.18 (br, -CH2-CH(2)O- (backbone), and -O-CH2-CH3 (side
chain)), 1.25−0.9 (-O-CH2-CH3 and -CH-CH3).
Removal of the Acetal Protective Groups. The PEEGE-containing
(co)polymer was dissolved in methanol (ca. 20 wt %) and the same
volume of 1 N HCl was added. The mixture was stirred at room
temperature overnight, concentrated in vacuo and precipitated three
times into diethyl ether. The isolated yields ranged between 80 and
1
90% in all cases. H NMR (DMSO-d6, 300 MHz): 10.57 (s, 1H, H-
N+R3), 7.54−6.97 (8H, aromatic signals from the initiator), 4.27 (br s,
OH, intensity and chemical shift can vary), 3.75 (s, 6H, MeO-C6H4-),
3.64−3.18 (br, -CH2-CH(2)O- (backbone)).
Hypergrafting of Glycidol. The linear precursor polymer was dis-
solved (or suspended) in benzene at a concentration of about 20 wt %.
Cesium hydroxide monohydrate was added (to achieve a degree of
deprotonation of 20%) and the mixture was allowed to react at 60 °C
for 60 min. The formed water and benzene were removed azeotro-
pically at 90 °C over a period of 90 min. After that, the activated
macroinitiator was suspended in dry diglyme (ca. 20 wt %), heated to
90 °C, and freshly distilled glycidol (amount depends on the targeted
degree of polymerization) in dry diglyme was added slowly with a
syringe pump over a period of 5−8 h. The reaction was terminated by
the addition of 20 mL methanol and 0.5 g acidic ion-exchange resin,
filtered, concentrated in vacuo, and precipitated into a 10-fold excess
of diethyl ether to afford the desired polymer in quantitative yield.
Hydrogenation (General Procedure). A hydrogenation vessel was
charged with the polymer (ca. 1 g) dissolved in degassed methanol (ca.
20 mL) and palladium on activated charcoal (ca. 100−200 mg) was
added under a stream of argon. The vessel was closed, pressurized with
hydrogen (8 bar) and stirred at room temperature until no residual
aromatic signals could be detected by 1H NMR spectroscopy (typically
24−48 h were necessary to achieve complete conversion). After
completion of the reaction, workup was done by purging the reaction
vessel with argon and filtration over Celite to remove the catalyst and
washed with about 100 mL of methanol. The filtrate was then
concentrated to about 5 mL, precipitated into cold diethyl ether and
dried. 1H NMR (DMSO-d6, 300 MHz): 5.0−3.80 (br, OH, intensity and
chemical shift can vary), 3.64−3.10 (br, -CH2-CH(2)O- (backbone)).
General Procedure for the Polymer Chain End Modification with
Squaric Acid Diethyl Ester. The polymer (0.05 mmol) was dissolved
in ethanol (15 mL) and pyridine was added (ca. 20 mg). To this
mixture a 5-fold molar excess of squaric acid diethyl ester (SADE, 0.25
mmol, 43 mg) was added with a microliter syringe and the mixture was
gently shaken at room temperature over a period of 16 h. The solution
Procedures. N,N-Di(p-methoxybenzyl)aminoethanol (1).
Freshly distilled p-methoxybenzyl bromide (7.5 g, 37.5 mmol),
ethanolamine (1.15 g, 18.5 mmol), and potassium carbonate
(7 g, 50 mmol) were mixed in about 80 mL of DMF and refluxed
for 24 h. After the reaction mixture was allowed to cool to room
temperature, the solution was filtered and diethyl ether (ca. 200 mL)
was added. The organic phase was then washed with water and a
saturated NaHCO3 solution and dried with MgSO4. The organic
phase was dried and concentrated in vacuo to afford a highly viscous
liquid. The crude product was purified by column chromatography
using ethyl acetate and petrol ether (6:4) as eluent. Yield: 4.5 g
1
(80%). H NMR (300 MHz, DMSO): δ (ppm) 7.26−6.85 (8H,
aromatic), 4.33 (t, J = 6 Hz 1H, OH), 3.72 (s, 6H, OCH3), 3.47 (s,
4H, NCH2Ph), 3.44 (t, J = 6 Hz, 2H, CH2OH), 2.65 (t, J = 6 Hz,
2H, NCH2). ESI-MS: 302.5 (MH+), 324.8 (MNa+).
N,N-Di(p-methoxybenzyl)serinol (2). Compound 2 was synthe-
sized as above with the following ratios: serinol (1.13 g, 12.4 mmol),
1163
dx.doi.org/10.1021/bm300103u | Biomacromolecules 2012, 13, 1161−1171

Biomacromolecules
Article
Scheme 2. Synthesis of Different Monoamino Polyethers Based on Polyethylene Glycol (PEG) and Polyglycerol (PG) with
a
Varying Architecture
a
(i) 1. CsOH·H2O, vacuo; 2. ethoxyethyl glycidyl ether; 3. HCl/MeOH. (ii) 1. CsOH·H2O, vacuo; 2. ethylene oxide and ethoxyethyl glycidyl ether;
3. HCl/MeOH. (iii) 1. CsOH·H2O, vacuo; 2. ethylene oxide; 3. ethoxyethyl glycidyl ether; 4. HCl/MeOH.
a
Scheme 3. Synthesis of Mid-Functional Poly(ethylene glycol-co-glycerol)s Starting from Di(p-methoxybenzyl)serinol (2)
a
EO: ethylene oxide; EEGE: ethoxyethyl glycidyl ether.
was concentrated in vacuo (taking care to keep the temperature below
60 °C to prevent transesterification) and precipitated four times into
diethyl ether. Typical yields are between 85 and 95%. 1H NMR
(DMSO-d6, 300 MHz): 8.7−8.4 (1H H-N-, two amide signals), 4.6
(br, 2H, −O-CH2-CH3), 3.64−3.18 (br, -CH2-CH(2)O- (backbone)),
1.4 (br, 3H, -O-CH2-CH3).
Lysozyme Activity Measurements. The activity of the lysozyme
conjugates was tested using Micrococcus lysodeikticus (Ml) cells as
substrate. Micrococcus lysodeikticus cells (17 mg) were suspended in
sodium acetate buffer (45 mL, pH = 5, 50 mM) and an aliquot of
3 mL was transferred into a UV cuvette. The initial absorbance at
450 nm was defined as baseline. Subsequently, 5 μL of a 1 g/L (protein)
solution of the appropriate polymer−protein conjugate in borate buffer/
DMSO (3:1) was added, the resulting solution mixed, and the
absorption at 450 nm was recorded as a function of time during a period
of 3 min. The activity of the conjugate is reported as a per-
centage compared to unmodified lysozyme with activity of 100%. The
reported activities are the average of three independent experiments.
General Procedure for Squaric Acid Mediated Protein Con-
jugation. BSA or lysozyme was dissolved in a 3:1 mixture of borate
buffer (0.01 M, pH = 9.1) and DMSO to yield a final protein
concentration of 1 g/L. The calculated amount of polymer (for BSA, 9
equiv, or for lysozyme, 1−7 equiv) was added to the protein solution
and the reaction mixture was gently shaken over a period of 20 h at
room temperature. Samples were taken from the reaction mixture and
dialyzed against deionized water (molecular weight cutoff of 1000 g/mol)
to remove salts and DMSO and were analyzed by SDS-PAGE to
monitor the conjugation reaction. Purification of the remaining solu-
tion was achieved by dialysis against deionized water for 48 h (depending
on the molecular weight of the polymer, dialysis membranes with
molecular weight cutoff between 15000 and 100000 g/mol were used
for lysozyme conjugates, while all BSA conjugates were dialyzed
against a molecular weight cutoff of 50000 g/mol), and subsequently
freeze-dried. Yield: 95% quantitative (compared to protein).
RESULTS AND DISCUSSION
■
Polymer Synthesis. A library of different poly(glycerol)
(PG) and poly(ethylene glycol-co-glycidol) copolymers was syn-
thesized via anionic polymerization as outlined in Schemes 2 and 3.
The library consists of linear PG and linear poly(ethylene glycol-
co-glycerol) copolymers (series A), midfunctional PEG-co-PG co-
polymers (series B), hyperbranched PG (series C), and linear-block-
hyperbranched PEG-b-hbPG copolymers (series D; Scheme 1).
1164
dx.doi.org/10.1021/bm300103u | Biomacromolecules 2012, 13, 1161−1171

Biomacromolecules
Article
Di(p-methoxybenzyl)aminoethanol (1) was used as the
initiator for the synthesis of linear, linear-hyperbranched, and
hyperbranched polymers,16,34 while di(p-methoxybenzyl)-
serinol (2) was used as initiator for the preparation of midfunc-
tional polymers. In the random copolymers the glycidol content
was adjusted to 10%. The molecular weights were varied by the
monomer: initiator ratio and several well-defined polymers
(Mw/Mn = 1.05−1.15 for the linear polymers, respectively,
1.20−1.50 for the hyperbranched polymers) with variation in
architecture and molecular weight were obtained. For the syn-
thesis of linear PG and poly(ethylene glycol-co-glycerol)
copolymers, ethoxy ethyl glycidyl ether32 was used as the
monomer, which was deprotected after polymerization by
diluted hydrochloric acid to release the hydroxyl groups
(see Supporting Information, Figure S1 for a representative
MALDI ToF mass spectrum). Hyperbranched PG was syn-
thesized via a slow-monomer addition hypergrafting protocol
onto linear PG as reported previously.16,35 For all synthesized
polymers, the di(p-methoxybenzyl) protected amino group was
released in a final step by catalytic hydrogenation at a hydrogen
pressure of about 8 bar with palladium on activated charcoal as
a catalyst. The end point of the reaction was determined by 1H
NMR spectroscopy monitoring the disappearance of the
aromatic signals of the protective group. After completion of
the reaction, all polymers showed a positive ninhydrin test. The
resulting amino-functionalized PG (co)polymers were charac-
terized by 1H NMR spectroscopy, SEC and MALDI ToF mass
Table 1. Molecular Weight Characteristics of the
Poly(glycerol) (Co)Polymers Investigated in this Study
c
d
e
e
Mn,th
Mn,NMR
Mn,SEC
Mw/Mn
(−)
a
b
series
A
sample code
lin-PG35
lin-PG74
(g/mol)
(g/mol)
(g/mol)
2500
5000
2000
5000
10000
20000
2000
5000
2000
4000
10000
20000
2500
5000
8000
15000
3000
4500
5500
2600
5500
2000
4500
10000
19400
2200
5500
2000
3400
9800
18000
2600
4800
6400
15000
2800
3300
5200
1900
5200
2200
3400
7600
14800
1700
4600
1800
3000
6500
13700
2500
4400
6000
13200
2710
2900
5000
1.12
1.18
1.05
1.05
1.07
1.12
1.09
1.11
1.04
1.09
1.12
1.15
1.30
1.28
1.52
1.39
1.26
1.18
1.20
lin-P(EG41-co-G3)
lin-P(EG102-co-G7)
lin-P(EG204-co-G14)
lin-P(EG407-co-G28)
mid-PG30
B
mid-PG68
mid-P(EG41-co-G3)
mid-P(EG100-co-G8)
mid-P(EG205-co-G14)
mid-P(EG367-co-G24)
hb-PG35
C
D
hb-PG65
hb-PG86
hb-PG203
PEG13-b-hbPG28
PEG52-b-hbPG14
PEG50-b-hbPG40
a
b
Compare Scheme 1. Subscripts indicate the number-average degree
c
of polymerization determined via 1H NMR. Theoretical number-
average molecular weight calculated from the initial monomer/initiator
ratio. Experimentally determined number-average molecular weight
via H NMR. Experimentally determined number-average molecular
weight (Mn) and polydispersity (Mw/Mn) via SEC in DMF vs poly-
(ethylene oxide) standards.
d
e
1
1
spectrometry. H NMR spectra of several selected polymers
are included in the Supporting Information (Figure S2−S5).
Table 1 lists the molecular weights of the different polymers as
1
determined from H NMR spectroscopy and SEC. Generally,
corresponding ammonium salts are formed, pyridine was added
for neutralization during the SADE modification, which may
facilitate dimerization.36 When 5 equiv of SADE were used, the
MALDI-ToF spectra indicated that no coupling takes places for
polymers with a molecular weight of 2000 g/mol and higher
and that quantitative end-capping is guaranteed. Low molecular
weight mPEG17 (Mn = 800 g/mol) showed a slight tendency
for dimerization (ca. 10% as estimated from MALDI ToF mass
spectrometry; see Supporting Information, Figures S6 and S7).
However, because they do not participate in the protein
modification reaction, dimerized polymer impurities are uncritical
and easily removed via dialysis after the conjugation reaction.
Modification of the amino-functionalized PG (co)polymers
using the conditions established above proceeded quantitatively
as indicated by a negative ninhydrin test. The resulting SADE-
modified PG (co)polymers were further characterized with
MALDI-ToF mass spectrometry and 1H NMR spectroscopy. As a
typical example, Figures 1 and 2 show the MALDI-ToF mass
the experimentally determined molecular weights were in good
agreement with those expected based on the monomer:
initiator ratio, which illustrates the living nature of the anionic
ring-opening polymerization.
Synthesis of Protein-Reactive Poly(glycerol) (Co)-
Polymers. To allow protein modification, the PG (co)polymers
shown in Schemes 2 and 3 need to be further modified with an
appropriate protein reactive functional group. The specific aim of
this study was to explore the squaric acid mediated coupling
strategy for the synthesis of PG (co)polymer−protein conjugates.
The synthesis of the protein-reactive, squaric acid modified PG
(co)polymers is outlined in Scheme 4.
In a first series of test reactions, the conditions for the
reaction outlined in Scheme 4 were optimized with respect to
the molar excess of SADE to minimize the formation of the
bisamide analogue. These experiments were carried out with
commercially available monoamino poly(ethylene glycol)s with
molecular weights of 750, 2000, 5000, and 10000 g/mol. In
general, dimerization by the formation of the diamide is un-
likely because the reactivity of the diester is much higher com-
pared to the ester amide. However, because during the
deprotection of the acetal functionalized prepolymers the
1
spectrum and the H NMR spectrum, respectively, of a midfunc-
tional poly(ethylene glycol-co-glycerol) copolymer (mid-P(EG100
-
co-G)8). The MALDI-ToF mass spectrum shows the typical
distribution of a copolymer with all possible linear combinations of
Scheme 4. Polymer Modification with Squaric Acid Diethyl Ester to Form a Protein-Reactive Polymer-Squaric Ethylester-Amide
(Polymer-SEA)
1165
dx.doi.org/10.1021/bm300103u | Biomacromolecules 2012, 13, 1161−1171

Biomacromolecules
Article
Figure 1. MALDI-ToF mass spectrum of a midfunctional random copolymer of PEG and PG (midP(EG100-co-G8)) modified with squaric acid ethyl
ester (Mn = 3400 g/mol (NMR)). The mass indicated in the insert at 2812 corresponds to the sum formula of P(EG50-co-G5).
contains 59 lysine residues38 of which 30−35 are accessible for
polymer modification.39 In a first set of experiments, several
BSA−polymer conjugates were prepared by dissolving the
protein in a 3:1 (v/v) mixture of borate buffer (ca. 1−3 g/L,
pH = 9.1, 10 mM) and DMSO and adding an 8-fold excess of
the SAE-activated polymer (Scheme 5). The mixture was gently
shaken overnight at room temperature and the reaction product
purified via dialysis.
The BSA−PG conjugates were first characterized by SEC.
The results of these analyses are summarized in Table 2. The
number of conjugated polymer chains per protein was
calculated from RALS detection after determination of the
apparent dn/dc and by considering the known molecular
weight of BSA and the conjugated PG (co)polymer. In addition
to the PG conjugates, Table 2 also lists the characteristics of
Figure 2. 1H NMR spectrum of a squaric acid activated midfunctional
two PEG conjugates that were obtained by reacting BSA with a
random copolymer of PEG and PG (mid-P(EG100-co-G8)).
10-fold excess of squaric acid monoamide functionalized PEG.
For the lower molecular weight P(E)G (co)polymers, Table 2
both monomer masses (44 for ethylene glycol and 74 for glycerol).
reveals excellent agreement between the expected and
The inset in Figure 1 only reveals a single distribution of species,
experimentally determined degree of P(E)Gylation. Increasing
which can be assigned to the SADE-modified polymer.
the molecular weight of the synthetic polymer generally results
in a decrease in the experimentally determined degree of
The 1H NMR spectrum in Figure 2 reveals the characteristic
squaric acid ethyl ester resonances at 1.37 (-CH3) and 4.66 ppm
P(E)Gylation as compared to the expected value. Similar find-
(-CH2-), respectively (note: the methylene resonance at 4.66 ppm
ings have been reported previously for other systems and were
is not visible in every sample due to possible overlap with the OH
resonances of the polyol), as well as signals due to the polymer
backbone between about 3.3 and 3.8 ppm and an amide resonance
at 8.6 and 8.8 ppm. The splitting of the amide signal into two
resonances has been reported before for low molecular weight
squaric acid derivatives and is a result of the rotational barrier
across the rigid cyclobutene carbon nitrogen bond and the
different chemical environment of the proton provided by the
syn rotamer and the corresponding anti form.37
ascribed to decreasing end group reactivity and steric crowding
at the protein with increasing molecular weight of the polymer.13
The SEC experiments further revealed that the Mark−Houwink
α-parameter of the PG-protein conjugates was influenced by
the polymer architecture. All conjugates with linear polymers
showed a value of about α = 0.33, which is an often found value
for a star-branched polymer,40 while conjugates prepared with
midfunctional or hyperbranched polymers showed lower α
values in the range of 0.19−0.24, that is, are more branched.
The lowest α values (0.11−0.13) are observed for the
hbPG and linear-hyperbranched block copolymer conjugates.
Protein Conjugation. BSA was selected as a model protein
to explore the squaric acid mediated polymer conjugation. BSA
1166
dx.doi.org/10.1021/bm300103u | Biomacromolecules 2012, 13, 1161−1171

Biomacromolecules
Article
Scheme 5. Conjugation of Squaric Acid Modified PG (Co)Polymers to Lysine Residues of a Protein
Table 2. Characteristics of the BSA−Polymer Conjugates
Prepared in This Study
a
a
Mn
dn/dc
a
a
b
c
sample code
BSA
BSA-(PEG44)n
BSA-(PEG113
(g/mol) Mw/Mn (−) (mL/g)
α (−) Nth NSEC
64400
87000
112250
1.08
1.17
1.12
1.11
0.16
0.14
0.13
0.14
0.16
0.36
0.38
0.29
0
0
10 9.8
10 9.3
)
n
BSA-(lin-P(EG41- 83100
co-G3)n
8
8
8
8
8
8
8.6
BSA-(lin-P(EG102
co-G7)n
-
-
-
106000
123000
158 000
97 300
114 000
1.22
1.13
1.10
1.08
1.11
0.12
0.10
0.10
0.12
0.12
0.32
0.33
0.33
0.24
0.19
8
Figure 3. SDS-PAGE (12%) of BSA−polyether conjugates: (Lane 1)
molecular weight markers; (Lane 2) BSA; (Lane 3) BSA-(lin-P(EG41-
co-G3)n); (Lane 4) BSA-(lin-P(EG102-co-G7)n); (Lane 5) BSA-(lin-
P(EG204-co-G14)n); (Lane 6) BSA-(lin-P(EG407-co-G28)n); (Lane 7)
BSA-(mid-P(EG41-co-G3)n); (Lane 8) BSA-(mid-P(EG205-co-G14)n);
(Lane 9) BSA-(hbPG35)n; (Lane 10) BSA-(PEG13-b-hbPG28)n.
BSA-(lin-P(EG204
co-G14)n
6
BSA-(lin-P(EG407
co-G28)n
4.6
6.3
5
BSA-(mid-
P(EG100-co-G8)n
BSA-(mid-
P(EG205-co-G14)n
and lysozyme during the conjugation reaction. Lysozyme
contains seven amine groups that are available for polymer
modification: six lysine residues and the N-terminal amine
group. To investigate to which extent the degree of polymer
modification can be controlled, first a series of screening experi-
ments with squaric acid activated α-methoxy-ω-amino poly-
(ethylene glycol)s (mPEG) with molecular weights of 750,
2000, 5000, and 10000 g/mol was carried out. In these experi-
ments, varying equivalents of squaric acid activated mPEG were
reacted with lysozyme at pH 9.1 in a 9:1 (v/v) borate buffer:
DMSO mixture and the degree of PEGylation investigated via
MALDI ToF mass spectrometry and electrophoresis (see
Supporting Information, Figures S8−S11). In all cases, MALDI
ToF mass spectrometry revealed a mixture of conjugates with
different degrees of PEGylation. Up to a targeted degree of
PEGylation of 4 and using PEG derivatives with molecular
weights of 750 or 2000 g/mol, the MALDI ToF mass spectra
suggest relatively good control over the number of conjugated
PEG chains. Increasing the mPEG/lysozyme ratio from 4:1 to
5:1 for a squaric acid activated mPEG derivative with a mol-
ecular weight of 2000 g/mol, in contrast, was not found to result
in a significant further improvement in the degree of PEGylation
(see Figure S9). Further increasing the PEG molecular weight to
5000 g/mol made it more difficult to generate lysozyme−polymer
conjugates where the degree of PEGylation corresponds to
the used stoichiometry of amine reactive PEG and lysozyme
(Figure S10).
BSA-(hbPG35)n
83 000
88 700
1.09
1.10
0.11
0.13
0.11
0.13
8
8
6.8
8.1
BSA-(PEG13-b-
hbPG28)n
a
Number average molecular weight (Mn), polydispersity index
(Mw/Mn), refractive index increment (dn/dc), and Mark−Houwink
parameter (α) determined by triple-detection SEC in water
(phosphate buffer, pH = 6.5, 0.1 M)/methanol (9:1). Expected
number of polymer chains conjugated per protein. Average number of
b
c
polymer chains per protein as detected from SEC.
The dn/dc values of the conjugates are approaching those of
the pure polymers with increasing mass percentage of the
polymer in the conjugate (note: the dn/dc for hbPG was found
to be 0.11 cm3/g under these conditions and is close to the
literature value of 0.12 in aqueous NaNO317).
To further corroborate the GPC analyses, the BSA
conjugates were also analyzed by gel electrophoresis. Figure 3
summarizes the results of these experiments. The gels in Figure
3 were run using unpurified conjugation mixtures, which were
only dialyzed to remove excess salt and DMSO before electro-
phoresis. The results of these experiments confirm the absence
of free, unmodified BSA, which underlines the efficiency of the
conjugation reaction under the investigated reaction conditions.
Furthermore, the gel electrophoresis experiments indicate that
the apparent molecular weight of the conjugates shifts to higher
values with increasing molecular weight of the conjugated
synthetic polymer.
Based on the results of the screening experiments discussed
above, a series of polyglycerol (co)polymer−lysozyme conjugates
were prepared with a targeted degree of modification of 3. As an
example, Figure 4 shows the MALDI-ToF mass spectrum of
lysozyme and the corresponding midfunctional P(EG100-co-G8)
conjugate. The mass spectrum confirms the successful
conjugation and reveals multiple distributions, which are due to
multiple charged and multimeric species as is expected for protein
analysis by MALDI ToF mass spectrometry.
In a second series of conjugation experiments, the enzyme
lysozyme was chosen as substrate for the modification with the
PG−(co)polymers. Lysozyme was selected as it is a convenient
model protein to assess the influence of PG-conjugation on
biological activity (vide infra). In contrast to the BSA modi-
fication experiments discussed above, which were carried out
with a large excess of PG, the objective of the lysozyme modi-
fication experiments was to control the degree of modification
by adjusting the relative quantity of squaric acid activated PG
1167
dx.doi.org/10.1021/bm300103u | Biomacromolecules 2012, 13, 1161−1171

Biomacromolecules
Article
and demonstrate the success of the conjugation reaction in all
cases. In most cases a small amount of free protein can be
detected in the unpurified conjugation mixture together with a
distribution of conjugates with a variable number of polymers
attached to the protein. If a 3-fold molar excess of two
architecturally different copolymers with similar molecular
weights is used (cf. lin-P(EG204-co-G14; Figure 5A, Lane 5) and
mid-P(EG205-co-G14) (Figure 5A, Lane 10)), one can clearly see
that the apparent molecular weights of the resulting conjugates
determined via SDS-PAGE are very similar. Further increasing
the molecular weight of the synthetic polymer (see, e.g., lin-
P(EG407-co-G28), Lane 6) results in a shift to higher apparent
molecular weights.
In addition to lysozyme conjugates in which only a part of
the available amine groups was modified, another series of
conjugation experiments was performed that was aimed at com-
plete modification of the lysozyme amine groups. Figure 6
shows SDS-PAGE analysis of lysozyme that was that was
reacted with 10 or 20 equiv of a linear (Lane 3), midfunctional
(Lane 2), or linear-hyperbranched polyglycerol derivative
(Lane 1) of a molecular weight of about 2000−2700 g/mol
as well as different hyperbranched polyglycerol−lysozyme con-
jugates (Lanes 4−7). As expected for the complete modi-
fication of lysozyme, SDS analysis of the hbPG35 conjugate
(Lane 7) reveals a single broad band. In contrast, for hbPGs
with higher molecular weight, that is, 4800, 6400, and 15000 g/
mol, conjugation is observed (which is detected by SDS-PAGE
(Figure 6, Lanes 4−6)) but the resulting conjugates show a
broader distribution of molecular weights. This may be due to
steric reasons as the protein-reactive SA ester amide is located
at the core of the hyperbranched polymer and probably less
accessible compared to linear polymers, so that only low
Figure 4. MALDI-ToF mass spectra of lysozyme (red) and a
conjugate of lysozyme and midfunctional P(EG100-co-G8) (blue).
Figure 5 shows the results of the SDS-PAGE analysis of a
series of lysozyme-polyglycerol (co)polymer conjugates that
were prepared using a 3-fold molar excess of the protein-
reactive polymer relative to lysozyme. These experiments were
carried out with samples taken from the crude, unpurified
reaction mixture. Figure 5A,B show the results for different
lysozyme conjugates prepared with linear and midfunctional
polyglycerol (co)polymers at a polymer/lysozyme ratio of 3:1
Figure 5. SDS PAGE of (A): (Lane 1) molecular weight markers; (Lane 2) lysozyme; (Lane 3) lin-P(EG41-co-G3) (2000 g/mol, 3 equiv); (Lane 4)
lin-P(EG102-co-G7) (4500 g/mol, 3 equiv); (Lane 5) lin-P(EG204-co-G14) (10000 g/mol, 3 equiv); (Lane 6) lin-P(EG407-co-G28) (19400 g/mol, 3
equiv); (Lane 7) linPG35 (2600 g/mol, 2 equiv); (Lane 8) linPG74 (5500 g/mol, 2 equiv); (Lane 9) molecular weight markers; and (Lane 10) mid-
P(EG205-co-G14) (9800 g/mol, 3 equiv); (B): (Lane 1) lin-P(EG204-co-G14) (10000 g/mol, 3 equiv); (Lane 2) molecular weight markers; (Lane 3)
mid-P(EG41-co-G3) (2000 g/mol, 3 equiv); (Lane 4) mid-P(EG100-co-G8) (3400/mol, 3 equiv); (Lane 5) mid-P(EG205-co-G14) (9800 g/mol,
3 equiv); (Lane 6) mid-P(EG367-co-G24) (18000 g/mol, 3 equiv); (Lane 7) mid-PG68 (5500 g/mol, 3 equiv); (Lane 8) mid-PG30 (2200 g/mol,
3 equiv); (Lane 9) lysozyme; and (Lane 10) molecular weight markers.
1168
dx.doi.org/10.1021/bm300103u | Biomacromolecules 2012, 13, 1161−1171

Biomacromolecules
Article
Table 3. Comparison of the Activities of Different Lysozyme
Conjugates
a
b
c
conjugated polymer
Mn,polymer (g/mol)
Npolymer
activity
PEG17
800
3
5
3
5
3
5
3
3
3
3
3
3
3
3
3
3
3
3
3
3
1
2
3
5
99 3.0
18 0.6
71 2.0
2.8 0.4
41 1.0
0.5 0.4
25 1.9
71 2.4
62 1.0
28 2.3
23 0.9
67 2.5
52 2.1
78 2.0
67 1.7
57 2.0
34 1.7
79 1.1
58 1.2
51 1.3
93 1.0
85 2.0
60 1.0
PEG17
800
PEG44
2100
2100
5100
5100
10100
2000
4500
10000
19400
2600
5500
2000
3400
9800
18000
2700
3300
5200
2500
2500
2500
2500
PEG44
PEG113
PEG113
PEG226
Figure 6. SDS-PAGE analysis of lysozyme conjugates prepared with
(Lane 1) PEG13-b-hbPG28 (2800 g/mol, 20 equiv); (Lane 2) mid-
P(EG41-co-G3) (2000 g/mol, 20 equiv); (Lane 3) lin-P(EG41-co-G3)
(2000 g/mol, 20 equiv); (Lane 4) hbPG203 (15000 g/mol, 10 equiv);
(Lane 5) hbPG86 (6400 g/mol, 10 equiv); (Lane 6) hbPG65 (4800 g/mol,
10 equiv); (Lane 7) hbPG35 (2500 g/mol, 10 equiv); (Lane 8) lysozyme;
(Lane 9) molecular weight markers.
lin-P(EG41-co-G3)
lin-P(EG102-co-G7)
lin-P(EG204-co-G14)
lin-P(EG407-co-G28)
lin-PG35
lin-PG74
mid-P(EG41-co-G3)
mid-P(EG100-co-G8)
mid-P(EG205-co-G14)
mid-P(EG367-co-G24)
PEG13-b-hbPG28
PEG52-b-hbPG14
PEG50-b-hbPG40
hbPG35
molecular weight polymers react with lysozyme while the larger
ones do not react. Conjugation of hbPG203 results in quite
sharp bands on the SDS PAGE, which probably represent
lysozyme was modified with one and two polymer chains (Lane 4)
as the main product. In addition several other conjugates can be
detected even with very high molecular weight, which do not
penetrate into the gel.
hbPG35
Enzymatic Activity. The activity of a series of PEG and PG
(co)polymer-lysozyme conjugates was investigated against
Micrococcus lysodeikticus (Ml) cells.41 Prior to the activity mea-
surements, all conjugates were purified by exhaustive dialysis to
remove unmodified lysozyme. Gel electrophoresis, HPLC and
GPC analysis results of the purified conjugates are included in
the Supporting Information (Figures S12−S17). The activity of
the conjugates was measured in acetate buffer at pH = 5 (50 mM)
as lysozyme has its highest activity at pH = 5.42 The results are
summarized in Table 3.
All polymer−lysozyme conjugates show a loss of activity as
compared to unmodified lysozyme as it has been reported in
literature for other lysozyme−polymer conjugates.13,42 From
the activity data, three general conclusions can be drawn: (i)
the more polymer chains (of a given molecular weight) are
conjugated to lysozyme the more the activity is decreased; (ii)
at any given degree of modification, the activity of the conjugates
(of the same polymer-type) decreases as the molecular weight of
the conjugated polymer increases; (iii) only at high PG
(co)polymer molecular weights, lysozyme activity appears to be
significantly influenced by polymer architecture.
To illustrate the first conclusion, if five PEG17 chains (Mn =
800 g/mol) were conjugated, the residual activity was found to
be 18 2%. For lysozyme modified with 5 equiv PEG44 (Mn =
2100 g/mol), the activity was reduced further to 3.8 1.2%
and essentially no activity was observed when in average five
chains PEG113 (Mn = 5100 g/mol) were conjugated to
lysozyme (residual activity 0.5 1%). When all amine groups
of lysozyme were P(E)Gylated (use of 20-fold molar excess of
the polymer), no residual activity was detected for most
conjugates (data not shown), which is also in agreement with
other studies that have been reported on various polymers.11
The influence of the degree modification on enzymatic activity
is also obvious from the results on the hbPG35 conjugates. The
activity of these conjugates decreases from 93 to 31% upon
increasing the degree of P(E)Gylation from 1 to 5.
hbPG35
hbPG35
31 5.0
a
1
Number average molecular weight (Mn) determined via H NMR.
b
c
Ratio polymer/protein used in the reaction. Activity measured
against Micrococcus lysodeikticus cells in percent compared to
unmodified lysozyme.
results obtained on the lin-P(EG-co-G) and mid-P(EG-co-G)
lysozyme conjugates. For both families of conjugates, the enzy-
matic activity was found to decrease from 71% (lin-P(EG-co-G)),
respectively, 78% (mid-P(EG-co-G)) to 23% (lin-P(EG-co-G)),
and 34% (mid-P(EG-co-G)) upon increasing the molecular weight
of the conjugated polymer from 2000 to ∼20000 g/mol.
In several studies, the influence of polymer architecture on
the biological activity of polymer−protein conjugates has been
reported. Tao et al., for example, reported a 2.5 times higher
activity of a polymer−protein conjugate consisting of a mid-
functional PHPMA (28.5% residual activity) as compared to
the linear counterpart (11.5%) of comparable molecular weight
(ca. 9000 g/mol).12 However, the compared protein conjugates
differed in the number of attached polymer chains (1.6 for the
midfunctional vs 3.1 for the linear), which makes an archi-
tectural comparison difficult. Another example was given for
uricase, which was PEGylated with a 10000 g/mol linear or
midfunctional PEG by Veronese et al.7 The authors found a
residual activity of 29% for linear PEG and 70% for midfunc-
tional PEG. On a first sight, comparison of the enzymatic
activities of the conjugates listed in Table 3 that have identical
degrees of modification but are generated from architecturally
different P(E)G copolymers of comparable molecular weight
does not provide any very obvious hints as to possible structure−
activity relationships. For example, at a degree of modification of 3
and a polymer molecular weight of ≈2000 g/mol, the PEG44 con-
jugate shows a relative activity of 71% as compared to 71, 78, 79,
and 60% for the comparable lin-P(EG41-co-G3), mid-P(EG41-co-
G3), PEG13-b-hbPG28, and hbPG35 conjugates. At a similar degree
The influence of polymer molecular weight at a given degree
of P(E)Gylation on the activity is most apparent from the
1169
dx.doi.org/10.1021/bm300103u | Biomacromolecules 2012, 13, 1161−1171

Biomacromolecules
Article
of modification but using polymers with a molecular weight of
≈5000 g/mol, the relative activities of the PEG113, lin-P(EG102-co-
G7), and PEG50-b-hbPG40 were determined as 41, 62, and 51%.
While possible effects of the P(E)G architecture on the activity of
the low molecular weight polymer conjugates seem to be relatively
minor, comparison of the activities of the higher molecular weight
P(E)G conjugates does reveal some interesting insights. From the
three conjugates that contain three copies of a P(E)G copolymer
with a molecular weight of ∼10000 g/mol, the residual activity of
the mid-P(EG205-co-G14) conjugate is two times higher than those
of the PEG226 and lin-P(EG204-G14) conjugates. Also, comparison
of the lin-P(EG407-co-G28) and mid-P(EG367-co-G24) conjugates,
which consists of synthetic polymers of comparable molecular
weights, shows a 1.5 times higher residual activity of the mid-
functional as compared to the linear P(E)G copolymer con-
jugate. These observations are in agreement with earlier studies
(vide supra)7,12 and indicate that, at similar molecular weights,
the use of a midfunctional, water-soluble polymer is advanta-
geous to minimize the loss of enzymatic activity as compared to
the use of the corresponding linear polymer.
as to minimize loss of enzymatic activity in the resulting modified
proteins.
ASSOCIATED CONTENT
■
S
* Supporting Information
Additional NMR, mass spectrometry, and SDS PAGE analysis.
This material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: harm-anton.klok@epfl.ch.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
F.W. is grateful to the Alexander von Humboldt Stiftung for a
Feodor Lynen fellowship. C.D. is a recipient of a fellowship of
the Max Planck Graduate Center mit der Johannes Gutenberg-
Universitat Mainz (MPGC). This work was partially supported
by the NCCR Nanoscale Science (H.-A.K.)
̈
CONCLUSIONS
■
This contribution has successfully demonstrated the feasibility
of the squaric acid dialkyl ester mediated amine selective modi-
fication of proteins with amine end functionalized, side chain
functional synthetic polymers. The synthetic polymers investi-
gated in this study were amine end-functionalized polyglycerol
(PG) copolymers obtained via ring-opening polymerization of
glycidol derivatives and ethylene oxide. The anionic ring-open-
ing polymerization of glycidol is a versatile synthetic strategy
that was used to prepare a library of linear, midfunctional, hyper-
branched, and linear-block-hyperbranched PG copolymers.
Whereas the efficiency of traditional approaches to couple
such hydroxyl side chain functional polymers to proteins can be
compromised by concurrent transesterification and hydrolysis
reactions, the squaric acid diethylester strategy presented herein
combines a high amine selectivity with low susceptibility to
hydrolysis, which enables the amine selective protein con-
jugation of side chain hydroxyl functionalized PG copolymers.
For the experiments in this study, BSA and lysozyme were used
as model proteins. Varying the relative amounts of squaric acid
diethylester activated polymer to protein allowed to control the
degree of modification of these proteins. The library of PG−
lysozyme conjugates was used to investigate the possible effects
of the architecture of the PG copolymers on the enzymatic
activity of the PG−lysozyme conjugates. Comparison of the
activity of PG−lysozyme conjugates generated from relatively
low molecular weight PG copolymers (<5000 g/mol) did not
reveal any obvious structure−activity relationships. Evaluation
of the activity of conjugates composed of PG copolymers with
molecular weights of 10000 or 20000 g/mol, however,
indicated significantly higher activities of conjugates prepared
from midfunctional synthetic polymers as compared to linear
polymers of similar molecular weight. The squaric acid diethyl
ester mediated protein conjugation presented here allows un-
precedented access to side chain functional synthetic polymer−
protein hybrid conjugates and may prove particularly useful for
the generation of polymer−protein hybrids modified with label-
ing groups or targeting ligands for biological or medical applica-
tions. Combined with the versatility of the glycidol ring-opening
polymerization, this unique coupling strategy opens the way to a
broad architectural variety of polymer−protein conjugates and
allows to engineer the architecture of the synthetic polymer such
REFERENCES
■
(1) James, E. T. Adv. Drug Delivery Rev. 1993, 10, 247−299.
(2) Putney, S. D.; Burke, P. A. Nat. Biotechnol. 1998, 16, 153−157.
(3) Veronese, F. M.; Pasut, G. Drug Discovery Today 2005, 10, 1451−
1458.
(4) Caliceti, P.; Veronese, F. M. Adv. Drug Delivery Rev. 2003, 55,
1261−1277.
(5) Abuchowski, A.; van Es, T.; Palczuk, N. C.; Davis, F. F. J. Biol.
Chem. 1977, 252, 3578−3581.
(6) Abuchowski, A.; McCoy, J. R.; Palczuk, N. C.; Es, T. v.; Davis, F. F.
J. Biol. Chem. 1977, 252, 3582−3586.
(7) Veronese, F. M.; Caliceti, P.; Schiavon, O. J. Bioact. Compat.
Polym. 1997, 12, 196−207.
(8) Pasut, G.; Veronese, F. M. Prog. Polym. Sci. 2007, 32, 933−961.
(9) Alconcel, S. N. S.; Baas, A. S.; Maynard, H. D. Polym. Chem.
2011, 2, 1442−1448.
(10) Bailon, P.; Palleroni, A.; Schaffer, C. A.; Spence, C. L.; Fung, W.-J.;
Porter, J. E.; Ehrlich, G. K.; Pan, W.; Xu, Z.-X.; Modi, M. W.; Farid, A.;
Berthold, W.; Graves, M. Bioconjugate Chem. 2001, 12, 195−202.
(11) Gauthier, M. A.; Klok, H.-A. Polym. Chem. 2010, 1, 1352−1373.
(12) Tao, L.; Xu, J.; Gell, D.; Davis, T. P. Macromolecules 2010, 43,
3721−3727.
(13) Tao, L.; Liu, J.; Davis, T. P. Biomacromolecules 2009, 10, 2847−
2851.
(14) Broyer, R. M.; Grover, G. N.; Maynard, H. D. Chem. Commun.
2011, 2212−2226.
(15) Gauthier, M. A.; Klok, H.-A. Chem. Commun. 2008, 2591−2611.
(16) Wurm, F.; Klos, J.; Rader, H. J.; Frey, H. J. Am. Chem. Soc. 2009,
̈
131, 7954−7955.
(17) Kainthan, R. K.; Brooks, D. E. Biomaterials 2007, 28, 4779−
4787.
(18) Wilms, D.; Stiriba, S.-E.; Frey, H. Acc. Chem. Res. 2009, 43,
129−141.
(19) Siegers, C.; Biesalski, M.; Haag, R. Chem.Eur. J. 2004, 10,
2831−2838.
(20) Dworak, A.; Walach, W.; Trzebicka, B. Macromol. Chem. Phys.
1995, 196, 1963−1970.
(21) Kainthan, R. K.; Muliawan, E. B.; Hatzikiriakos, S. G.; Brooks,
D. E. Macromolecules 2006, 39, 7708−7717.
(22) Sunder, A.; Hanselmann, R.; Frey, H.; Mulhaupt, R.
Macromolecules 1999, 32, 4240−4246.
(23) Roberts, M. J.; Bentley, M. D.; Harris, J. M. Adv. Drug Delivery
Rev. 2002, 54, 459−476.
1170
dx.doi.org/10.1021/bm300103u | Biomacromolecules 2012, 13, 1161−1171

Biomacromolecules
Article
(24) Greenwald, R. B.; Pendri, A.; Martinez, A.; Gilbert, C.; Bradley, P.
Bioconjugate Chem. 1996, 7, 638−641.
(25) Subr, V.; Ulbrich, K. React. Funct. Polym. 2006, 66, 1525−1538.
́
(26) Hou, S.-j.; Saksena, R.; Kovac, P. Carbohydr. Res. 2008, 343,
196−210.
(27) Izumi, M.; Okumura, S.; Yuasa, H.; Hashimoto, H. J. Carbohydr.
Chem. 2003, 22, 317−329.
(28) Yan, H.; Aguilar, A. L.; Zhao, Y. Bioorg. Med. Chem. Lett. 2007,
17, 6535−6538.
(29) Sejwal, P.; Han, Y.; Shah, A.; Luk, Y.-Y. Org. Lett. 2007, 9,
4897−4900.
(30) Tietze, L. F.; Arlt, M.; Beller, M.; Glusenkamp, K. H.; Jahde, E.;
̈
̈
Rajewsky, M. F. Chem. Ber. 1991, 124, 1215−1221.
(31) Dingels, C.; Wurm, F.; Klok, H.-A.; Frey, H. submitted for
publication.
(32) Fitton, A. O.; Hill, J.; Jane, D. E.; Millar, R. Synthesis 1987, 1140,
1142.
(33) Zhang, W.; Go, M. L. Eur. J. Med. Chem. 2007, 42, 841−850.
(34) Mangold, C.; Wurm, F.; Obermeier, B.; Frey, H. Macromol.
Rapid Commun. 2010, 31, 258−264.
(35) Wurm, F.; Nieberle, J.; Frey, H. Macromolecules 2008, 41,
1184−1188.
(36) Schmidt, A. H. Synthesis 1980, 961−994.
(37) Rotger, M. C.; Pina, M. N.; Frontera, A.; Martorell, G.; Ballester, P.;
̀
Deya, P. M.; Costa, A. J. Org. Chem. 2004, 69, 2302−2308.
(38) Hirayama, K.; Akashi, S.; Furuya, M.; Fukuhara, K.-I. Biochem.
Biophys. Res. Commun. 1990, 173, 639−646.
́
(39) Van Regenmortel, M. H. V., Briand, J. P., Muller, S.; Plaue, S.
̈
Synthetic Polypeptides as Antigens; Elsevier: Amsterdam, 1988; Vol. 19.
(40) Held, D.; Muller, A. H. E. Macromol. Symp. 2000, 157, 225−
̈
238.
(41) Salton, M. R. J. Nature 1952, 170, 746−747.
(42) Nodake, Y.; Yamasaki, N. Biosci. Biotechnol. Biochem. 2000, 64,
767−774.
1171
dx.doi.org/10.1021/bm300103u | Biomacromolecules 2012, 13, 1161−1171