The molecular recognition of phosphorylated proteins by designed polypeptides conjugated to a small molecule that binds phosphate

Article abstract of DOI:10.1039/c1ob06154b

The conjugation of polypeptides from a designed set to the small molecule ligand 3,5-bis[[bis(2-pyridylmethyl)amino]methyl]benzoic acid, which in the presence of Zn²⁺ ions binds inorganic phosphate, has been shown to provide a polypeptide conju

Full text of DOI:10.1039/c1ob06154b

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 7697
www.rsc.org/obc
PAPER
The molecular recognition of phosphorylated proteins by designed
polypeptides conjugated to a small molecule that binds phosphate†
´
Adam T. Slo´sarczyk and Lars Baltzer*
Received 12th July 2011, Accepted 31st August 2011
DOI: 10.1039/c1ob06154b
The conjugation of polypeptides from a designed set to the small molecule ligand 3,5-bis[[bis(2-
pyridylmethyl)amino]methyl]benzoic acid, which in the presence of Zn2+ ions binds inorganic
phosphate, has been shown to provide a polypeptide conjugate that binds a-casein, a multiply
phosphorylated protein, with a dissociation constant KD of 17 nM. The measured affinity is more than
three orders of magnitude higher than that of the small molecule ligand for phosphate and the binding
of 500 nM of a-casein was not inhibited by 10 mM phosphate buffer, providing a 2000-fold excess of
phosphate ion over protein. The selectivity for phosphoproteins was demonstrated by extraction of
a-casein from solutions of various complexity, including milk and human serum spiked with a-casein.
In addition to a-casein, b-casein was also recognized but not ovoalbumin. Conjugation of a
polypeptide to the zinc chelating ligand was therefore shown to give rise to dramatically increased
affinity and also increased selectivity. A set of polypeptide conjugates is expected to be able to capture a
large number of phosphorylated proteins, perhaps all, and in combination with electrophoresis or mass
spectrometry become a powerful tool for the monitoring of phosphorylation levels. The presented
binder can easily be attached to various types of surfaces; here demonstrated for the case of polystyrene
particles. The example of phosphoproteins was selected since posttranslational phosphorylation is of
fundamental importance in cell biology due to its role in signaling and therefore of great interest in
drug development. The reported concept for binder development is, however, quite general and
high-affinity binders can conveniently be developed for a variety of proteins including those with
posttranslational modifications for which small molecule recognition elements are available.
would otherwise be considered poor or failed, or at an early stage
Introduction
of development. In particular, conjugation to a polypeptide has
been shown to induce a considerable difference in affinity for
proteins to which the small molecule binds with equal strength,
suggesting that specificity can be substantially improved.2
Small organic molecules that recognize and bind proteins are
attractive goals in pharmaceutical and bioanalytical chemistry,
yet notoriously difficult to develop to the highest levels of
performance. New molecular strategies that enable us, with less
effort, to increase affinities and selectivity for protein targets
are therefore of considerable interest. Recently we reported a
novel concept for protein recognition and binding, based on the
conjugation of small organic molecules to polypeptides where
the resulting conjugates were shown to have affinities several
orders of magnitude higher than those of the small molecules
and to have improved selectivity.1–3 Binder molecules based on
this concept are readily modified by conjugation techniques
and readily immobilized on solid support for chip-based and
continuous flow applications.
While high-affinity binders for specific proteins are extremely
important, binder molecules for specific functional groups that
arise, for example, as a result of post-translational modifications
are also of considerable interest. The concept behind polypeptide
conjugate binder molecules is that the small molecule warhead
dominates the interactions between binder and protein, and that
the attached polypeptide enhances affinity as well as selectivity.1
Under the condition that the small molecule warhead is selective
for a specific functional group rather than a specific protein
surface epitope or active site, the resulting binder molecules are
expected to be capable of recognizing large fractions of the proteins
carrying such functional groups. A panel of binders could be
developed that bind all proteins modified to carry that particular
functional group, in a biological sample. This application is
especially important in proteomics in the search for proteins
that are not known previously to undergo posttranslational
modification, where antibody based detection cannot be used.
In this context, the extractions of groups of proteins, followed by
The concept provides an opportunity to form specific, high-
affinity binders for proteins taking advantage of compounds that
Department of Biochemistry and Organic Chemistry, Uppsala University,
Box 576, 751 23, Uppsala, Sweden. E-mail: lars.baltzer@biorg.uu.se;
Fax: +46 18-471-3818; Tel: +46 18-471 3810
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c1ob06154b
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7697–7704 | 7697
©

View Article Online
mass spectrometric analysis or pull-down experiments followed
by Western blots are attractive goals. In a first application of this
strategy we have addressed the recognition of proteins modified
by phosphorylation.
in HEPES buffer (pH = 7.2, 10 mM HEPES, 150 mM NaCl),
the second one contained 500 nM of 4C15*L8-PP1 and 500 nM
a-casein in HEPES buffer (pH = 7.2, 10 mM HEPES, 150 mM
NaCl), the third one contained 500 nM of 4C15*L8-PP1 and
1000 nM a-casein in HEPES buffer (pH = 7.2, 10 mM HEPES,
150 mM NaCl). Each solution was prepared immediately prior
to measurement. Each series of the three solutions was measured
in triplicate and the mean value of the three measurements was
reported. The fluorescence emission of the solution containing
only protein dissolved in the buffer was registered but the response
was only at the noise level (not reported).
Phosphorylation is a reversible covalent modification controlled
by kinases and phosphatases and acts as a “molecular trigger”
for signal transduction cascades, apoptotic progression, metabolic
changes and gene expression.4 Whenever abnormal phosphoryla-
tion occurs, mutagenic, neuropathogenic or cancerogenic activities
are initiated and kinases are therefore important drug targets. The
interest in phosphotyrosines from a pharmaceutical perspective
stems largely from their involvement in cell signaling and therefore
in cancer.5 Phosphoserines and phosphothreonines are approx-
imately 2000-fold more abundant and their phosphorylation
occurs downstream from that of tyrosines.6 The relatively poor
performance of anti-phosphoserine and anti-phosphothreonine
antibodies suggest that there is a need for high-affinity, selective,
robust and efficient binders for proteins phosphorylated at the side
chains of serines and threonines.
Phospho-specific antibodies are routinely used to recognize
phosphorylated epitopes in immunoprecipitation7–9 and although
antiphosphotyrosine antibodies show satisfactory efficiency10,11
antiphosphoserine and antiphosphothreonine antibodies are as
a rule only moderately specific.12 Non-specific capturing of
phosphate groups is used in combination with mass spectrometry,
using immobilized metal ion chromatography (IMAC) or metal
oxides such as TiO2.12–14 Upon modification or elimination of
phosphate groups in proteins, chemical tags, such as biotin, can
be introduced for extraction and identification of phopshorylated
proteins.15 The latter methods typically require that the proteins
are digested by trypsin before capture and mass spectrometric
analysis.
Affinity determination by fluorescence spectroscopy
A series of solutions were prepared, where each one contained
the binder 4C15L8-PP1* (fluorescein attached to the cysteine at
position 24) at a concentration of 100 nM and the concentration
of a-casein was varied from 1nM to 10 mM. All solutions were
prepared in HEPES buffer (pH = 7.2, 10 mM HEPES, 150 mM
NaCl). The fluorescence at 525 nm was determined as a function
of total protein concentration and the dissociation constant KD
was determined by fitting eqn (1) to the experimental results.
Fbound[aC]+ Ffree KD
F
=
(1)
obs
[aC]+KD
In eqn (1), Fobs is the observed fluorescence intensity, Fbound is the
fluorescence of the peptide-binder bound to the a-casein, Ffree is
the fluorescence of the free peptide and [aC] is the concentration
of the free a-casein. [aC] can be derived from eqn (2), where
[P]tot is the total concentration of peptide and [aC]tot is the total
concentration of the a-casein.
2
Here, we have addressed the problem of recognition and
binding of phosphorylated proteins, and in a proof-of-principle
demonstration applied the binder concept to the capturing of the
model proteins a-casein, b-casein and ovalbumin.16,17 Both caseins
carry multiple phosphate groups that may enhance the probability
for binding and thus their avidities. The binder molecules have
been evaluated with regards to affinity, binding mechanism and
selectivity in media of variable complexity. The results suggest
that these binder molecules are strong candidates for diagnostic
applications in the cases where phosphorylation events play vital
roles.
⎛
⎜
⎜
⎞
[P]tot +KD −[aC]tot
[P]tot +KD −[aC]tot
⎟
[aC]=−
+
+KD[aC]tot
⎟
⎟
⎟
⎜
⎝
⎠
2
2
(2)
Numerical fitting was done with IGOR Pro 4.03 (WaveMetrics
Inc.).
Pull-down experiments
Polystyrene beads (0.5 mg–1 mg) coated and functionalized
with polypeptide conjugate binder molecules as described in the
supporting information were incubated with buffer, milk or serum
(0.5 mL–1 mL). The suspensions were shaken end-over-end for
90 min and centrifuged (14 000 rpm for 7 min). The resulting
bead plug was washed at least 4 times by means of sequential
resuspension in 1 mL of HEPES buffer (pH = 7.2, 10 mM HEPES,
150 mM NaCl), followed by centrifugation (14,000 rpm for 7
min). In between the washing procedure an occasional sonication
was conducted in order to diminish unspecific binding to the
surface. After the washing procedure, the beads were centrifuged
and the plug was resuspended in the DTT solution (1 mM) in
HEPES buffer (pH = 7.2, 10 mM HEPES, 150 mM NaCl) (20 mL)
and incubated for 1 h with occasional sonication. Thereafter, the
suspension was centrifuged and the supernatant was analysed by
SDS-PAGE and stained using Silver Staining. Samples for gel
electrophoresis were prepared according to the manufacturer’s
instructions.
Experimental
A detailed description of the organic synthesis, the solid phase
peptide synthesis, the deprotection and purification of polypep-
tides as well as the conjugation of the small organic molecule
and fluorophor to the polypeptides is provided in the supporting
information.
Fluorescence screening
Fluorescence measurements were carried using standard 96 well
plates, coated to prevent non-specific binding as described in
the supporting information, and a microtitre plate reader. Three
solutions were prepared: the first one contained 500 nM of
4C15*L8-PP1 (coumarin attached to the lysine at position 15)
7698 | Org. Biomol. Chem., 2011, 9, 7697–7704
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
“Catch and release” column experiment
protein and Leu, Ile or Phe residues on the hydrophobic face
of the amphiphilic helices.3 The enhanced selectivity arises as
a consequence of electrostatic interactions between amino acid
residues in the polypeptide and those exposed on the protein
surface. The total charge of the peptides varies between -7 and
+2 and the majority of the charged residues were introduced close
to the hydrophobic face of each helix to bring them in close contact
with residues on the protein surface. In the present investigation
eight sequences were singled out for binder development, Fig. 2,
whereas in other cases the full set of 16 sequences was screened to
identify the best binders with the highest affinity and selectivity.1,2
Here the polypeptides with total charges of -1, the so called 3-
series, and +2, the so called 4-series, were used.
The polystyrene beads were coated and functionalized with binder
molecules as described above. Functionalized beads (3 mg) were
applied on the Biosphere Filter Tips as a suspension in HEPES
buffer (pH = 7.2, 10 mM HEPES, 150 mM NaCl). The beads were
washed with HEPES buffer containing 150 mM NaCl at pH 7.2
and 300 mL of the analyte was applied to the Filter Tips. The beads
were washed with 50 mL of the HEPES buffer four times and the
eluate from the fourth washing was collected for further analysis.
In order to release the bound proteins the beads were washed once
with 20 mL of 0.5% TFA and the eluate was collected. Both the
fourth washing and the eluate were analysed by SDS-PAGE.
Results and discussion
Binder design and synthesis
In general, the binder concept is based on a set of 16 designed
polypeptide sequences1–3 and for each target protein a selective
small molecule warhead is selected and covalently linked to each
member of the set, Fig. 1. Depending on the target a spacer may
be incorporated to link the small molecule to the polypeptides.
Fig. 2 Amino acid sequences in one-letter code of eight polypeptides used
for binder development. Lysine residues used for conjugation to PP1 are
shown in bold and lysine residues used for conjugation of coumarin are
shown in italics.
By design each peptide was equipped with three orthogonal
anchoring positions that could be addressed independently, two
lysine residues and one cysteine, Fig. 2. Selective modification
of the lysines was achieved using orthogonal protection groups.
The cysteine is located in the loop region and is conveniently
used for immobilization on beads or attachment of further
functional groups. Modifications can be performed both on the
solid phase during peptide synthesis and in solution after cleavage
and purification. An illustration of the folded structure and its
interactions with a protein is shown in Fig. 1.
Among many available molecules that bind the phosphate
group21 a Zn(II)–2,2¢-dipicolylamine (DPA) complex, reported
previously by Hamachi et al., to bind the phosphate anion was
used.22,23 We selected a molecule composed of two DPA groups
linked by dimethylbenzoic acid for binder development, Fig. 1.
It has been used previously in the construction of phosphate
receptors and binds two Zn2+ ions that coordinate to the phosphate
oxygens. The complex has been reported to bind phosphopeptides
with dissociation constants in the mM range.24 The 3,5-bis[[bis(2-
pyridylmethyl)amino]methyl]benzoyl group, PP1, was attached to
the polypeptide by amide formation to the side chain of a lysine
residue.
The synthesis of 3,5-bis[[bis(2-pyridylmethyl)amino]methyl]-
benzoic acid is outlined in Scheme 1.25 The synthesis route was
optimized, see supporting information, and the target compound
produced in very good yield compared to that reported in the
literature.17
Fig. 1 Illustration of the binder concept and the interactions between the
polypeptide conjugate binder and a schematic phosphorylated protein in
the presence of Zn2+. No crystal structure of a-casein exists and the protein
is therefore presented in a schematic way. Hydrophobic and charge–charge
interactions between polypeptide and the phosphorylated protein give rise
to affinity and selectivity that is considerably increased in comparison to
that between the small molecule and the protein. The Cys residue in the
loop was used for immobilization on beads.
The polypeptides were designed to have some propensity for
the formation of amphiphilic helices connected by a short loop.1,3
They dimerize to form molten globule like four-helix bundles at
mM concentrations but dissociate to unordered monomers at nM
concentrations.18 The polypeptides were not designed to form
preorganized structures but to be able adapt to the protein surface
interactions between the polypeptide and the target protein in
addition to those between the small molecule and the protein.1,2,19,20
The major part of the binding energy between protein and peptide
is believed to arise from hydrophobic interactions between the
The binder molecules were formed in a reaction between 5 and
the lysine side chain of the peptide attached to the solid support
using PyBOP (benzotriazol-1-yloxy)tripyrrolidinophosphonium
hexafluorophosphate) to ensure a high yield. The binder was
cleaved from the resin and purified by HPLC.
For affinity determination, the binder molecules were equipped
with fluorophores. For the initial studies, 7-methoxycoumarin-3-
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7697–7704 | 7699
©

View Article Online
Binder affinity
In a proof-of-concept demonstration we studied the binding of the
model proteins a-casein, b-casein and ovalbumin, where a-casein
is probably the most widely investigated and well characterized
phosphoprotein. It has between 8 and 13 phosphate groups
whereas b-casein has 5 and ovalbumin one. In these proteins
phosphorylation occurs at the side chains of serine residues.
They are readily available from commercial sources and therefore
suitable for systematic investigations. Both caseins and ovalbumin
have isoelectric points of around 5, with an IEP of 4.6 for a-casein,
5.1 for b-casein, and 4.6 for ovalbumin.
For the determination of affinity, fluorescent probes were em-
ployed that were expected to respond with altered quantum yields
to changes in the environmental hydrophobicity resulting from
binding. In a preliminary and simple assay executable in microtitre
plate format, each of the eight binder candidates were titrated with
a-casein in three steps (Supporting Information SM Fig. 1 and
SM Fig. 2†). To peptide stock solutions of approximately 5 mM
concentration as determined from the absorbance of the coumarin
probe at 345 nm, was added Zn(NO3)2 to a concentration of
12 mM. Aliquots from stock solutions were transferred to the
wells of microtitre plates and the final concentrations of binders
were 500 nM in 10 mM HEPES buffer containing 150 mM NaCl
at pH 7.2. Each of the binder molecules were equipped with
the fluorophor 7-methoxycoumarin-3-carboxylic acid, Fig. 2, and
the fluorescence emission spectra were recorded in the absence
of protein, as well as in the presence of 500 nM, 1000 nM and
1500 nM protein, Fig. 3.
Scheme 1 Outline of the synthesis of 3,5-bis[[bis(2-pyridylmethyl)amino]-
methyl]benzoic acid triethylammonium salt.
carboxylic acid was conjugated to the lysine side chain through
the reaction of 7-methoxycoumarin-3-carboxylic acid pentafluo-
rophenyl ester (6) in DMSO (dimethyl sulfoxide)/pyridine (py)
solution. The synthesis of 6 is outlined in Scheme 2.
Scheme 2 Outline of synthesis of 7-methoxycoumarin-3-carboxylic acid
pentafluorophenyl ester.
For accurateaffinitydetermination using abinder concentration
of 100 nM, fluorescein-5-pyrrolidine-2,5-dione was conjugated to
the deprotected thiol side chain of the cysteine residue through the
reaction with the commercially available fluorescein-5-maleimide.
Phosphorylation takes place on the surface of folded proteins
and therefore no spacer was included, PP1 was conjugated to the
side chain of a lysine residue.
Structural characterization
The secondary structure content of the set of sixteen unmodified
polypeptides has been reported previously.1,2 In order to determine
whether the introduction of PP1 and Zn2+ ions affected the
solution conformation of the polypeptides the binder molecule
4C15L8-PP1 at various stages of assembly was investigated by
circular dichroism spectroscopy at 1, 10 and 50 mM concentration
in 10 mM HEPES (4-(2-hydroxyethyl)piperazine-1-ethanesulfonic
acid) buffer at pH 7.2 (Supporting Information: SM Table 1†).
All peptides showed helical signatures with minima at 209 and
222 nm suggesting that they are at least partly dimerised to form
four-helix bundles at these concentrations.18 The parent peptide 4-
C15L8 and the fully decorated 4-C15L8 carrying the Zn2+ complex
of PP1 showed a modest concentration dependence indicating
a monomer–dimer equilibrium whereas 4-C15L8 functionalized
with PP1 did not show concentration dependence in the absence of
Zn2+. We conclude that regardless of modification the polypeptides
fold into helix–loop–helix motifs that dimerize to four helix
bundles at concentrations in the mM range, and that their solution
structures are not strongly affected by the introduction of the
functional groups.
Fig. 3 Titration of 500 nM solutions of binder molecules with a-casein
in 10 mM HEPES buffer containing 150 mM NaCl at pH 7.2. A. The
fluorescence emission of the tight binder 4C15L8PP1(Zn2+
) decreases
2
upon addition of one equivalent of a-casein and is unaffected by further
additions. B. The emission increase of the weak binder 3C15L8PP1(Zn2+
)
2
does not reach saturation upon addition of one equivalent of a-casein
but increases with further additions. C. The fluorescence emission of the
polypeptide 4C15L8 is not affected by the presence of protein and does not
bind a-casein (lysine side chain acetylated). D. The fluorescence emission
of 4C15L8-PP1 in the absence of Zn2+ does not change showing that Zn2+
is required for binding.
The 4-series of binder molecules showed nearly identical fluores-
cence emission spectra regardless of the concentration of a-casein
suggesting that binding is saturated at 500 nM concentration
each of binder and protein, Fig. 3. From the fluorescence
titrations, assuming that the experimental error is not higher
than 10%, the 4-series of binder candidates were all estimated
to have dissociation constants of around 5 nM or lower. The
7700 | Org. Biomol. Chem., 2011, 9, 7697–7704
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
3-series of binders with the possible exception of 3C10L17-PP1,
(Supporting Information, SM Fig. 2†) did not reach saturation
at the protein concentrations used here. We conclude that the 3-
series binders have dissociation constants in the high nM to the
mM range. The nearly identical affinities within the 3-series and
within the 4-series are probably fortuitous. As negative controls
the unmodified polypeptide 4C10L17 and 4C10L17-PP1 in the
absence of Zn2+ were titrated with a-casein. No effects on the flu-
orescence emission spectra were observed as a function of protein
concentration.
The affinity of free polypeptide for a-casein is clearly low
and from a comparison between the affinity of the polypeptide
conjugate and the published dissociation constant of PP1 one can
conclude that conjugation of the polypeptide to PP1 gives rise
to a binder molecule that binds on the order of three orders of
magnitude stronger than PP1 and consequently, that the polypep-
tide contributes significantly to binding when covalently linked
to PP1. For a more accurate determination of the dissociation
constant of 4C15L8-PP1 in the presence of Zn2+, titration with a-
casein in 10 mM HEPES buffer and 150 mM NaCl at pH 7.2 was
carried out, Fig. 4. At a binder concentration of 100 nM a more
strongly emitting probe was required and fluorescein was attached
to the side chain of Cys-24 by reacting the maleimide group of
fluorescein-5-maleimide with the free thiol of the polypeptide after
removal of the acetamidomethyl (Acm) protection group. The
binder concentration was kept constant at 100 nM and a series
of samples were prepared where the concentration of a-casein
was varied. The best fit of an equation describing the dissociation
of a 1 : 1 complex to the experimental results gave an apparent
dissociation constant of 17 nM, although the description of the
complex dissociation behaviour of a-casein with 8–13 phosphate
groups and one or more copies of 4C15L8-PP1 in terms of a
single dissociation constant is obviously an oversimplification. It
is not known, for example, how many binder molecules bind to
the phosphate groups of a-casein, and it is not known if they are
associated with different affinities, although most likely they are
since the electrostatic profile of the protein surface is expected to
vary between phosphorylated sites.
The measured value is probably an underestimate of the true
KD. In order to obtain the most reliable value of a dissociation
constant, the measurements need to be carried out at a concen-
tration that is similar to the value of the dissociation constant.
In this case it was not feasible since the changes in fluorescence
emission were not large enough to permit measurements at or
below 17 nM concentration with the required accuracy. Also
the concentration of a-casein is lower than the concentration of
4C15L8-PP1 at around the inflexion point and below, and the
change in fluorescence emission therefore underestimates binding
strength. Unfortunately due to practical limitations linked to
fluorophore emission, we were unable to establish more accurate
values of KD.
In order to rank the affinities of the binder molecules pull-
down experiments with a-casein were instead carried out using the
binders with the highest affinities. Polystyrene latex nanoparticles
were functionalized with26 all the members of the 4-series of
binders by reacting deprotected Cys side chains in the loop regions
ofthe polypeptides withactivated coated particles. Free thiols react
readily with pyridylsulfide (PDS) groups to form disulfide bridges.
Particles coated with Pluronic F-107 carrying PDS groups were
incubated with 10 nM and 100 nM solutions of a-casein, washed
three times with buffer and separated from the supernatant after
each wash by centrifugation. After three washes the particles were
treated with DL-dithiothreitol (DTT) to release the binders and
captured proteins from the beads and the resulting supernatants
were analyzed by SDS-PAGE (Supporting Information: SM Fig
3†). All 4-series peptide conjugates successfully extracted a-casein
at a concentration of 100 nM whereas only the binders 4C10L17-
PP1 and 4C15L8-PP1 extracted a-casein from solutions with a
protein concentration of 10 nM. These two binders were the most
potent, clearly due to the best shape and charge complementarity
to a-casein.
The interactions between the binder 4C15L8-PP1 and a-casein
in the presence of Zn2+ were further probed by competition
experiments and analysed by pull-down followed by SDS-PAGE,
Fig. 5. Although PP1 chelated with Zn2+ was designed to bind
phosphorylated amino acid side chains through coordination to
the phosphate group, 10 mM PBS buffer corresponding to an
excess of phosphate anions over phosphoserines by three orders
of magnitude, did not inhibit the capture of a-casein at a protein
concentration of 500 nM. This opens up a wide range of diagnostic
applications in media with high concentrations of phosphate
anions.27–29 The capture of a-casein from a 500 nM solution
by 4-C15L8-PP1 was not inhibited by a synthetic peptide with
one phosphotyrosine side chain, PhosPep, even at a PhosPep
concentration of 400 mM. Only when the concentration of a-
casein was decreased to 100 nM and the concentration of PhosPep
was 80 mM did binding appear to be suppressed. Commercially
available dephosphorylated a-casein with a phosphorylation level
of 20% was also readily extracted by the binder, see supporting
information.
Fig. 4 Fluorescence intensity versus total concentration of a-casein
(upper panel) and free concentration of a-casein (lower panel). A
dissociation constant of 17 nM was obtained from the best fit to the
experimental results of an equation describing the dissociation of a 1 : 1
complex.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7697–7704 | 7701
©

View Article Online
Fig. 6 SDS-PAGE analysis of proteins extracted by binder molecules
immobilized on beads coated with Pluronics from protein mixture
containing 500 nM lysozyme (15 kDa), 500 nM phosphorylase B (97
kDa), 500 nM b-galactosidase (116 kDa), 500 nM a-casein (23.6 kDa)
and 500 nM ovalbumin (45 kDa). Panel A) Extract by 4C15L8-PP1
Lane 1. Extract from 500 nM protein mixture. Lane 2. Extract from
500 nM a-casein by Pluronic coated beads, negative control. Lane 3.
500 nM protein mixture, positive control. Lane 4: Extract from 500 nM
a-casein by 4C15L8-PP1 in the absence of Zn2+, negative control. Lane 5.
Extract from 500 nM protein mixture by 4C15L8-PP1 in the absence of
Zn2+, negative control. Lane 6. Extract from protein mixture by 4C15L8
acetylated at Lys8, negative control. Weak bands at the bottom of gel is
binder molecule, MW 5 kD. Panel B) Extracts from 500 nM protein mixt.
by binder molecules. Lane 1. 4C10L17-PP1 Lane 2. 4C25L22-PP1 Lane 3.
4C37L34-PP1. Lane 4. Extract from protein mixture by beads coated with
Pluronic, negative control. Lane 5. Protein mixture (Lysozyme not visible),
positive control. Panel C) Extract from 100 nM solution of b-casein in 10
mM HEPES buffer, 150 mM NaCl at pH 7.2 Lane 1. b-casein, positive
control. Lane 2. Extract by 4C15L8-PP1. Lane 3. Extract by beads coated
with Pluronics, negative control.
Fig. 5 SDS-PAGE analysis of a-casein extraction by 4C15L8-PP1in
competition with phosphate anion in PBS (A) and in competition
with the peptide PhosPep where one phosphotyrosine residue has been
incorporated in the sequence, in 10 mM HEPES, 150 mM NaCl at pH
7.2. (B). Panel A. Lane 1. a-casein, positive control. Lane 2. Extract from
500 nM a-casein in 10 mM PBS. Panel B. Lane 1. a-casein, positive control.
Lane 2. Extract from 500 nM a-casein and 400 mM PhosPep. Lane 3.
Extract from 500 nM a-casein and 40 mM PhosPep. Lane 4. Extract from
100 nM a-casein and 80 mM PhosPep1. Lane 5. Extract from 100 nM
a-casein and 8 mM PhosPep. Lane 6. Extract from 500 nM a-casein using
beads coated with Pluronics without binder molecules, negative control.
Binder selectivity
The analysis of extracts from protein mixtures containing 500 nM
each of a-casein, ovalbumin and the non-phosphorylated proteins
lysozyme, phosphorylase B and b-galactosidase was undertaken
to investigate the level of selectivity that could be obtained, Fig. 6.
Three negative controls were used for detection of unspecific
binding, beads coated with Pluronic, beads coated with Pluronic
conjugated to 4C15L8 where the lysine at position 8 was acetylated
and beads coated with Pluronic conjugated to 4C15L8-PP1 and
used in the absence of Zn2+ ions. To make sure that Zn2+ ions were
removed from the complex the extraction was carried out in the
presence of 1 mM EDTA.
All 4-series binders were capable of extracting a-casein selec-
tively from the mixture, with insignificant non-specific uptake.
A band corresponding to the non-phosphorylated protein phos-
phorylase B was observed but a control experiment showed that
phosphorylase B binds non-specifically to the Pluronic coated
beads. It was also confirmed that Zn2+ is essential for a-casein
binding since no a-casein was extracted in the presence of 1 mM
EDTA. The binders 4C10L17-PP1 and 4C15L8-PP1 provided the
highest intensity a-casein bands in comparison with those of the
other binder molecules.
from a 100 nM solution was demonstrated, Fig. 6. This protein
has approximately half the number of phosphate groups of a-
casein. The mono-phosphorylated protein ovoalbumine at 100 nM
concentration could not be extracted at a detectable level under the
experimental conditions. The difference in extraction performance
may be due to differences in affinity, since the avidity for poly-
phosphorylated proteins is expected to be higher than for a mono-
phosphorylated one. It may also be due to less optimal interactions
between polypeptide and protein, and could possibly be improved
by selecting a different polypeptide.
The selectivity of 4C15L8-PP1 in more complex media was
evaluated by pull-down experiments in bovine milk and human
serum, Fig. 7. Bovine milk contains, in addition to many other
proteins and lipids, a number of casein isoforms where a-casein is
the most abundant.30 Human serum contains thousands of differ-
ent proteins and extraction from human serum provides a critical
evaluation of binder selectivity of relevance to the measurement
of phosphorylated proteins in biomedical applications. Human
serum was selected as the medium although it does not contain
To evaluate binding also of phosphoproteins with a lower
level of phosphorylation than a-casein the extraction of b-casein
7702 | Org. Biomol. Chem., 2011, 9, 7697–7704
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
beads with a higher mean diameter in comparison with the beads
used in pull-down experiments. The binder molecules 4C15L8-
PP1 and 4C10L17-PP1 were selected as they were shown to be
the most efficient in capturing a-casein in pull-down experiments.
A mixture containing dilute human serum enriched with a-casein
was passed through the columns followed by washing. The release
of the bound proteins was achieved by 0.5% aqueous TFA and
aliquots were collected and analysed by SDS-PAGE, Fig. 8.
Extraction of a-casein from human serum was demonstrated at
very low concentration. Although unspecific binding to the beads
that was observed in pull-down experiments was also observed
in this experiment, a clear and dominating a-casein band was
found on the gel. To demonstrate that a-casein was extracted
specifically by the binder molecule, the final washing solution was
analysed by SDS-PAGE, Fig. 8. No a-casein band was observed,
proving that the captured protein was released only by the change
of pH and that it was extracted exclusively by specific interactions
with the binder molecule. The experimental results open a wide
Fig. 7 Pull-down of a-casein from milk (A) and human serum (B).
Panel A). Extract from 10-fold diluted bovine milk. Lane 1. Control,
a-casein. Lane 2. Extract by 4-C15L8-PP1. Panel B. Extracts from 100-fold
diluted human serum spiked with 100 nM a-casein (lanes 2–5) and 5-fold
diluted human serum spiked with 100 nM a-casein (lanes 8–9). Lane
1. Positive control a-casein, Lane 2. Extract by 4C10L17-PP1. Lane
3. Extract by 4C15L8-PP1. Lane 4. Extract by, 4C25L22-PP1. Lane 5.
Extract by4C37L34-PP1. Lane 6. Extract by beads coated with Pluronic,
negative control. Lane 7. 100 nM a-casein in 100-fold diluted human
serum, positive control. Lane 8. Extract by 4C10L17-PP1. Lane 9. Extract
by 4C15L8-PP1, Lane 10. Extract by beads coated with Pluronic, negative
control. Lane 11. Five-fold diluted human serum spiked with 100 nM,
positive control.
a-casein, which was therefore added to a concentration of 100 nM.
Pull-down of a-casein from bovine milk was clearly demonstrated
and only a minor band corresponding to b-casein accompanied
the main a-casein band. Pull-down of a-casein by all four binders
of the 4-series from a 100 nM spiked solution in 100-fold diluted
human serum was successfully demonstrated and pull-down from
spiked neat human serum was successful although several bands
appeared on the gel. It is not possible to avoid non-specific uptake
by the beads in neat human serum and it is thus not possible
to clearly define the level of selectivity of 4-C15L8-PP1 in neat
serum. However, it was demonstrated that a-casein was extracted
in competition with thousands of proteins by the synthetic binders
of the 4-series in 5-fold diluted serum spiked with 100 nM a-casein.
Some non-specific uptake is observed but otherwise a remarkable
selectivity by 4C15L8-PP1 whereas the other two binders give rise
to somewhat lower intensity. This result is in good agreement with
observations reported above.
Fig. 8 SDS-PAGE analysis of fractions collected from columns packed
with polystyrene beads coated with Pluronic and binder molecules
4C15L8-PP1 and 4C10L17-PP1. Analysed solution contained 100 fold
diluted human serum (HS) spiked with 10 nM a-casein, Lane 1. Final
wash from 4C10L17-PP1 column. Lane 2. Eluate released with 0.5%
aqueous TFA from 4C10L17-PP1 column. Lane 3. Final wash from
4C15L8-PP1 column. Lane 4. Eluate released with 0.5% aqueous TFA
from 4C15L8-PP1 column. Lane 5. Extract by beads coated with Pluronic,
negative control. Lane 6. Extract by beads coated with Pluronic, negative
control. Lane 7. a-casein, positive control Lane 8. Human serum, positive
control Lane 9. Human serum spiked with 10 nM a-casein, positive
control. Lane 10. Molecular mass standard.
In order to evaluate the possibility to utilize this technology in
continuous flow systems we prepared a set of columns. These
columns were packed with the beads functionalized with our
binders in the analogous way to the ones that were used in pull-
down experiments. For technical reasons, we were forced to use
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 7697–7704 | 7703
©

View Article Online
range of new applications, especially in combination with mass
spectrometric detection.
of Wrocław is gratefully acknowledged for providing HR-MS
data.
Conclusion
Notes and references
1 L. T. Tegler, G. Nonglaton, F. Bu¨ttner, K. Caldwell, T. Christopeit, H.
Danielson, K. Fromell, T. Gossas, A. Larsson, P. Longati, T. Norberg,
R. Ramapanicker, J. Rydberg and L. Baltzer, Angew. Chem., Int. Ed.,
2011, 50, 1823–1827.
2 L. T. Tegler, K. Fromell, B.-H. Jonsson, J. Viljanen, C. Winander, J.
Carlsson and L. Baltzer, ChemBioChem, 2011, 12, 559–566.
3 L. Baltzer, Anal. Bioanal. Chem., 2011, 400, 1653–166.
4 F. Marks, (Ed), Protein Phosphorylation, 1996.
5 T. Hunter, Cell, 2000, 100, 113–127.
6 T. Hunter and J. A. Cooper, Cell, 1981, 24, 741–752.
7 M. J. Comb, A Scientific Newsletter from New England Biolabs, 1995,
7, 1.
8 J. P. Blaydes, B. Vojtesek, G. B. Bloomberg and T. R. Hupp, Methods
in Molecular Biology, 2000, 99, 177–189.
9 H. Zhang, X. Zha, Y. Tan and P. V. Hornbeck et al., J. Biol. Chem.,
2002, 277, 39379–39387.
10 J. Rush, A. Moritz, K. A. Lee and A. Guo et al., Nat. Biotechnol., 2005,
23, 94–101.
11 J. Villen, S. A. Beausoleil, S. A. Gerber and S. P. Gygi, Proc. Natl. Acad.
Sci. U. S. A., 2007, 104, 1488–1493.
12 D. T. McLachlin and B. T. Chait, Curr. Opin. Chem. Biol., 2001, 5,
591–602.
The binders reported here show high affinity and selectivity for the
model phosphoproteins a-casein and b-casein in buffer as well as
in milk and serum and provide strong support for the hypothesis
that the set of polypeptide sequences used to develop the binders
for phosphoproteins and previously for the C-reactive protein
and human Carbonic Anhydrase II is of general applicability in
developing binders for proteins. The concept is different from that
of biological binders such as antibodies and aptamers where highly
complex and preorganized structures with high molecular weights
are required and offer an alternative for capturing and detecting
proteins in solution as well as on solid support. This type of binder
is highly suitable for a wide range of applications in biotechnology
and biomedicine where robust capturing technologies can extend
shelf-lives and areas of application quite significantly. The binders
can be stored without special precautions at room temperature and
are readily derivatized using well established chemical methods
and reagents. The possibility of conjugating fluorescent probes
and radionuclide sites specifically makes the technology highly
attractive in bioanalytical applications. We believe that the binder
will be capable of acting as an imaging agent, in vitro and in
vivo, due to the flexibility with which organic molecules can
be covalently attached without perturbing binding performance.
The binder molecules reported here are developed especially
for the purpose of monitoring phosphorylation events involving
serine and threonine. Enrichment of the biological samples prior
to mass spectrometric analysis is an attractive application for
binders that capture phosphorylated proteins bearing in mind
that phosphorylated proteins are less ionizable in comparison
to the corresponding non-phosphorylated ones. The binder also
performs very well in “pull-down” experiments and is expected
to have a large number of applications in monitoring cellular
phosphorylation events as a result of exposure to pharmaceuticals.
The fact that the casein proteins are both multiply phosphorylated
may have contributed to the efficient extraction achieved by
immobilized binder molecules, if the loading is high. However, in
solution only minor avidity effects are expected as the polypeptides
do not dimerize at the low concentrations used for titrations.
13 H. K. Kweon and K. Hakansson, Anal. Chem., 2006, 78, 1743–
1749.
14 M. R. Larsen, T. E. Thingholm, O. N. Jensen, P. Roepstorff and T. J.
Jorgensen, Mol. Cell. Proteomics, 2005, 4, 873–886.
15 Y. Oda, T. Nagasu and B. T. Chait, Nat. Biotechnol., 2001, 19, 379–
382.
16 A. Ojida, Y. Mito-oka, K. Sada and I. Hamachi, J. Am. Chem. Soc.,
2004, 126, 2454–2463.
17 A. Mangalum and R. C. Smith, Tetrahedron, 2009, 65, 4298–
4303.
18 S. Olofsson, G. Johansson and L. Baltzer, J. Chem. Soc., Perkin Trans.
2, 1995, 2047.
19 L. Baltzer, Top. Curr. Chem., 2007, 277, 89–106.
20 K. Enander, G. T. Dolphin, B. Liedberg, I. Lundstro¨m and L. Baltzer,
Chem.–Eur. J., 2004, 10, 2375–2385.
21 S. Aoki and E. Kimura, Rev. Mol. Biotechnol., 2002, 90, 129–
155.
22 A. Ojida, M. Inoue, Y. Mito-oka, H. Tsutsumi, K. Sada and I. Hamachi,
J. Am. Chem. Soc., 2006, 128, 2052–2058.
23 Y. Ishida, M. Inoue, T. Inoue, A. Ojida and I. Hamachi, Chem.
Commun., 2009, 2848–2850.
24 S. Yamaguchi, I. Yoshimura, T. Kohira, S. Tamaru and I. Hamachi, J.
Am. Chem. Soc., 2005, 127, 11835–11841.
25 P. Liu, Y. Chen, J. Deng and Y. Tu, Synthesis, 2001, 2078–2080.
26 K. Fromell, M. Andersson, K. Elihn and K. D. Caldwell, Colloids Surf.,
B, 2005, 46, 84–91.
27 S. Rapoport and G. M. Guest, J. Biol. Chem., 1941, 138, 269–
282.
28 E. P. Lehman, J. Biol. Chem., 1921, 48, 293–303.
29 P. Gustin, B. Detry, M. L. Cao, F. Chenut, A. Robert, M. Ansay, A.
Frans and T. Clerbaux, J Appl Physiol., 1994, 77, 202–208.
30 http://www.foodsci.uoguelph.ca/dairyedu/chem.html.
Acknowledgements
We are indebted to the Swedish Research Council for financial
support. Dr Piotr Stefanowicz, Faculty of Chemistry, University
7704 | Org. Biomol. Chem., 2011, 9, 7697–7704
This journal is
The Royal Society of Chemistry 2011
©