Designed, folded polypeptide scaffolds that combine key biosensing events of recognition and reporting

Full text of DOI:10.1021/jo010954n

3120
J . Org. Chem. 2002, 67, 3120-3123
Design ed , F old ed P olyp ep tid e Sca ffold s
Th a t Com bin e Key Biosen sin g Even ts of
Recogn ition a n d Rep or tin g
Karin Enander,† Gunnar T. Dolphin,†
Linda K. Andersson,‡ Bo Liedberg,§
Ingemar Lundstro¨m,§ and Lars Baltzer*,†
Department of Chemistry-IFM and Department of Applied
Physics-IFM, Linko¨ping University, 581 83 Linko¨ping,
Sweden, and Department of Organic Chemistry,
Go¨teborg University, 412 96 Go¨teborg, Sweden
Lars.Baltzer@ifm.liu.se.
Received September 26, 2001
The de novo design of folded polypeptides aims at
improving our understanding of protein structure and
also provides a platform for the engineering of new
proteins with tailored functions.1-3 Designed, folded
polypeptides that undergo pH-controlled, site-selective
self-functionalization with ligands4,5 constitute an excel-
lent toolbox for the construction of various complex
molecular systems, e.g., model glycoproteins6,7 or complex
receptors. With a focus on the development of functional
units in biosensors, we present in this paper, folded,
ligand-modified helix-loop-helix polypeptide scaffolds
that connect the key biosensing events of recognition and
reporting. The well-characterized interaction between the
enzyme human carbonic anhydrase II (HCAII) and its
benzenesulfonamide inhibitor8 (4-carboxybenzenesulfona-
mide, Ia ) was selected for a proof of principle demonstra-
tion. However, the variety of molecules that can be
incorporated in the polypeptides and the ease by which
their relative positions can be varied allow for a system-
atic development of biosensing units for a wide range of
receptor-ligand systems.
The design of the peptides KE2 and KE3 (Figure 1)
was based on the sequence of LA-42b, a 42-residue
polypeptide that folds into a helix-loop-helix motif and
dimerizes to form a four-helix bundle in aqueous solu-
tion.6 Out of 42 residues, more than 32 were conserved
in the design of KE2 and KE3. The solution structure of
LA-42b has been extensively studied by NMR and CD
spectroscopy, and because of the sequence similarity with
KE2 and KE3, they were assumed to fold into helix-
loop-helix dimer motifs as well. KE2 and KE3 were
synthesized using solid-phase peptide synthesis and
F igu r e 1. Modeled structure of KE2 and KE3, showing sites
of introduction of dansyl (position 15) and benzenesulfonamide
(positions 34 and 8, respectively). Only the amino acid side
chains in positions involved in functionalization are shown and
only those in the sequence of KE2. Amino acid sequences of
KE2 and KE3, where lysine residues in bold represent sites
of modification, are also shown. The amino terminals of both
peptides were acetylated.
identified by mass spectrometry. The MALDI-TOF spec-
tra of KE2 and KE3 with Cys22 protecting group intact
showed single peaks at 4446.3 and 4563.2, respectively
(calcd 4446.0 and 4563.2). The peptides were designed
to allow for a site-specific incorporation of a fluorescent
probe at the side chain of Lys15, as well as of a ligand
with high affinity for a target protein at the side chains
of Lys34 (KE2) or Lys8 (KE3). The side chain of Lys15
was orthogonally protected to allow the coupling of a
fluorescent probe on the solid phase. Before cleaving of
the peptide from the resin, the Lys15 Alloc protection
group was removed by 3 equiv of Pd(PPh3)4 in a mixture
of 9.25 mL ofCHCl3, 0.5 mL of AcOH, and 0.25 mL of
morpholine. Reaction of the selectively deprotected pep-
tides with 2 equiv of dansyl chloride in the presence of 8
equiv of diisopropylethylamine in DMF provided KE2-P
and KE3-P. The MALDI-TOF spectra of KE2-P and
KE3-P showed single peaks at 4680.4 and 4868.2, re-
spectively (calcd 4679.4 and 4867.6). The notation -P
indicates that a fluorescent probe has been covalently
attached, and the notation -PL indicates the attachment
of both fluorescent probe and high-affinity ligand. In all
experiments Cys22 remained protected.
† Department of Chemistry-IFM, Linko¨ping University.
‡ Go¨teborg University.
§ Department of Applied Physics-IFM, Linko¨ping University.
(1) DeGrado, W. F.; Summa, C. M.; Pavone, V.; Nastri, F.; Lombardi,
A. Annu. Rev. Biochem. 1999, 68, 779-819.
(2) Micklatcher, C.; Chmielewski, J . Curr. Opin. Chem. Biol. 1999,
3, 724-729.
(3) Baltzer, L.; Nilsson, H.; Nilsson, J . Chem. Rev. 2001, 101, 3153-
3163.
(4) Broo, K.; Brive, L.; Lundh, A. C.; Ahlberg, P.; Baltzer, L. J . Am.
Chem. Soc. 1996, 118, 8172-8173.
(5) Baltzer, L.; Nilsson, J . Curr. Opin. Biotechnol. 2001, 12, 355-
The incorporation of the benzenesulfonamide ligand
was accomplished by reacting the polypeptides with the
active ester Id in aqueous solution at pH 8. We have
previously shown that, after Lys15, Orn34 is the most
reactive of all lysine and ornithine residues in LA-42b9
in terms of its reactivity toward active esters, because it
360.
(6) Andersson, L.; Stenhagen, G.; Baltzer, L. J . Org. Chem. 1998,
63, 1366-1367.
(7) Andersson, L. K.; Dolphin, G. T.; Kihlberg, J .; Baltzer, L. J .
Chem. Soc. Perkin Trans. 2 2000, 459-464.
(8) Eriksson, A. E.; Kylsten, P. M.; J ones, T. A.; Liljas, A. Proteins
1988, 4, 283-293.
10.1021/jo010954n CCC: $22.00 © 2002 American Chemical Society
Published on Web 04/02/2002

Notes
J . Org. Chem., Vol. 67, No. 9, 2002 3121
is situated in a position that forms part of the hydropho-
bic core and has a selectively depressed pKa value.10 Some
competition from Lys19 was observed9,10 and therefore
Lys19 was replaced by Arg19 in the sequence of KE2.
Lys10 was also replaced by an Arg residue. No competi-
tion from Lys33 was expected on the basis of the previous
investigation, and it was therefore not removed. In KE3,
where Lys8 was introduced, all competing lysines were
replaced by Arg residues.
The affinity of HCAII for unprotonated Ia in aqueous
phosphate buffer at room temperature and pH 6.5 has
been reported previously, and the equilibrium dissocia-
tion constant Kd was 27 µM.11 Preliminary affinity
measurements by surface plasmon resonance suggested
that Kd for KE2-PL modified with Ia was in the µM
range, and we reasoned that, in order to avoid competi-
tion from nonspecific binding of HCAII, a higher affinity
ligand than that corresponding to Ia was needed. With
benzenesulfonamide derivatives bearing alkyl chains of
different lengths in the para position, increased affinities
toward HCAII as compared to that of Ia have been
reported.12 Thus, to increase the affinity of the benzene-
sulfonamide inhibitor, and to minimize sterical con-
straints upon HCAII binding that could be introduced
by coupling the benzenesulfonamide ligand to the pep-
tide, an aliphatic spacer was introduced. The N-hydroxy-
succinimidyl ester of 4-carboxybenzenesulfonamide (Ib)
was reacted with 6-aminohexanoic acid to form Ic, which
was further activated with N-hydroxysuccinimide to form
Id . Id (1.4 equiv) was allowed to react with 1 equiv of
KE2-P or KE3-P in 50 mM Tris-HCl buffer at pH 8.0 and
room temperature. The modified peptides KE2-PL and
KE3-PL were purified by reversed-phase HPLC on a
Hichrom C-8-column using 0.1% TFA in 40% aqueous
2-propanol as the eluent and lyophilized. The yield was
55% and 77% for KE2-PL and KE3-PL, respectively. The
MALDI-TOF spectra of KE2-PL and KE3-PL showed
single peaks at 4977.2 and 5165.0, respectively (calcd
4975.7 and 5163.9).
The biosensing capabilities of KE2-PL and KE3-PL
were investigated by recording their fluorescence emis-
sion spectra between 450 and 650 nm upon excitation at
335 nm. Spectra were recorded of 1 µM peptide solutions
in HBS buffer (10 mM Hepes, 0.15 M NaCl, 3 mM EDTA,
0.005% polysorbate 20) at pH 7.4 and 298 K, in the
absence and presence of 50 µM HCAII, with 1 µM
solutions of KE2-P and KE3-P as negative controls
(Figure 2). Upon addition of HCAII, the fluorescence
intensities of KE2-PL and KE3-PL increased by 80% and
60%, respectively, whereas the fluorescence intensities
of the control peptides were not affected by HCAII. The
observed intensity increases were thus caused by the
binding of HCAII to the benzenesulfonamide moiety of
the peptide scaffolds, with negligible effects induced by
nonspecific protein-peptide interactions. We interpret
the intensity increases to arise from a change in the
molecular environment of the dansyl group upon binding
of HCAII by the polypeptides; it appears that the probe
is partially quenched in the unbound peptide, but to a
F igu r e 2. Fluorescence spectra of 1 µM KE2 (A) and KE3
(B) modified with the dansyl fluorescent probe. Bold and thin
curves correspond to peptides with and without benzene-
sulfonamide, respectively (KE2-PL, KE3-PL; KE2-P, KE3-P).
Dotted lines represent spectra in the absence of HCAII, and
solid lines represent spectra where 50 µM HCAII has been
added. The fluorescence of the control peptides KE2-P and
KE3-P is not affected by the presence of HCAII, but the
fluorescence of KE2-PL is increased by 80%. KE3-PL displays
a similar but less pronounced behavior.
lesser extent when bound to HCAII. The polypeptides
KE2-PL and KE3-PL are therefore capable of reporting
on the presence of the target protein HCAII. Recognition
is ensured by the specificity of the benzenesulfonamide
ligand.
Upon titration of 1 µM KE2-PL with 5 nM-50 µM
HCAII, a sigmoid curve was obtained (Figure 3). For the
bimolecular association between KE2-PL and HCAII, the
equilibrium dissociation constant Kd equals the concen-
tration of free HCAII at the inflection point. Kd was
estimated to be 0.02 µM from the best fit to the experi-
mental results of an equation describing the dissociation
of a bimolecular complex. This result constitutes the proof
of principle for functional helix-loop-helix-based bio-
sensor units, since binding results in fluorescence inten-
sity changes. The use of an array of peptides modified
with ligands of different affinities makes measurements
of analyte concentrations possible, at levels of accuracy
limited, in principle, only by the number of different
ligand variants available and by the affinity range of
those variants.
(9) Andersson, L. K.; Dolphin, G. T.; Baltzer, L. ChemBioChem In
press
(10) Andersson, L. K.; Caspersson, M.; Baltzer, L. Chem. Eur. J . In
press
(11) Taylor, P. W.; King, R. W.; Burgen, A. S. V. Biochemistry 1970,
The CD spectra of 1 µM solutions of KE2-P and KE3-P
in 10 mM phosphate buffer at pH 7.5 revealed a higher
degree of helical content than that of the template peptide
9, 2638-2645.
(12) King, R. W.; Burgen, A. S. V. Proc. R. Soc. London B 1976,
193, 107-125.

3122 J . Org. Chem., Vol. 67, No. 9, 2002
Notes
Sch em e 1
LA-42b at comparable concentrations.7 The mean residue
ellipticity [θ]222 was -19 100 ( 1000 and -17 300 ( 1000
deg cm2 dmol-1 for 1 µM KE2-P and KE3-P, respectively,
and that of 2 µM LA-42b was -8000 ( 1000 deg cm2
dmol-1. The large difference between the KE-P peptides
and LA-42b suggests that the probe has an effect on helix
stability, possibly due to interactions between the probe
and the hydrophobic core of the peptides, or due to the
removal of a positive charge upon probe attachment.7
Difference CD spectra obtained by subtracting that of 2
µM HCAII from that of 1 µM KE2-PL + 2 µM HCAII
showed that the helicity of KE2-PL was unchanged upon
binding to HCAII. At these concentrations, and based on
a dissociation constant of 0.02 µM, more than 90% of the
peptide is bound to HCAII.
There was a significant difference between the fluo-
rescence intensities of KE3-P and KE3-PL in the absence
of HCAII. The attachment of a ligand in position 8 close
to the probe in position 15 may change the peptide
structure enough to decrease the quenching. This hy-
pothesis is supported by the observed differences in
helical contents between 1 µM KE3-P and 1 µM KE3-
PL, [θ]222 ) -17 300 ( 1000 and -24 700 ( 1000 deg
cm2 dmol-1, respectively.
HCAII and KE2-PL in solution was monitored for a series
of peptide concentrations ranging from 40 nM to 10 µM.
The steady-state affinity was estimated to be 0.08 µM
from curve fitting to a plot of equilibrium responses as a
function of peptide concentration using a 1:1 binding
model. Slightly nonideal behavior was observed, which
might be explained by an influence on the affinity by
concentration-dependent supersecondary structure for-
mation of the peptide. The steady-state affinity of HCAII
for Ic was also determined and Kd was found to be 0.044
µM. The results suggest that the affinity of HCAII for
the ligand decreases due to steric effects when the ligand
is attached to the peptide. An even longer spacer might
be useful to increase the affinity of HCAII for the peptide
even further. Lower affinity was observed with the
surface plasmon resonance-based assay than with the
solution-based fluorescence method. This may reflect
steric repulsion and/or orientational restrictions in the
presentation of the enzyme active site when HCAII was
immobilized. Similar results have been reported previ-
ously for the benzenesulfonamide-carbonic anhydrase
system.13
Our results demonstrate the first steps in the success-
ful application of folded helix-loop-helix peptides to the
area of biosensing. A peptide that binds to a receptor has
been shown to be able to report on this event via a
fluorescent probe. Furthermore, this concept has been
applied to the determination of an affinity constant. The
possibility of conveniently incorporating a wide range of
probes and ligands at different relative positions provides
an attractive way of optimizing the biosensing conditions,
such as sensitivity and fluorescence response, for any
target biomacromolecule. As indicated by our results, the
structure of the peptide scaffold also plays an important
role in sensor performance. The molecular basis for
biosensing is now under investigation in our laboratory.
A determination of analyte concentration is possible
using an array of peptides modified with ligands of
different affinities. To simplify the use of such peptide-
based biosensor arrays, immobilization strategies for
coupling to regenerable surfaces will be developed.
The affinity obtained from fluorescence spectroscopy
was compared to that measured with surface plasmon
resonance, where the interaction between immobilized
Ack n ow led gm en t. K.E. is enrolled in the graduate
school Forum Scientium and in the research program
Biomimetic materials science, both financed by the
Swedish Foundation for Strategic Research (SSF). Fi-
F igu r e 3. Sigmoid binding curve obtained upon titration of
1 µM KE2-PL with HCAII by plotting the maximum fluores-
cence intensity versus the logarithm of free HCAII concentra-
tion. The affinity of the interaction was estimated to be 0.02
µM from curve fitting using a two-state binding model.
(13) Mrksich, M.; Grunwell, J . R.; Whitesides, G. M. J . Am. Chem.
Soc. 1995, 117, 12009-12010.

Notes
J . Org. Chem., Vol. 67, No. 9, 2002 3123
CD spectroscopy and surface plasmon resonance-based affinity
analysis. This material is available free of charge via the
Internet at http/::pubs.acs.org.
nancial support was also obtained from the Swedish
Natural Science Research Council (NFR) and A+ Science
Invest AB, Go¨teborg, Sweden.
Su p p or tin g In for m a tion Ava ila ble: Sequence of the
template peptide LA-42b and figures displaying results from
J O010954N