A versatile polypeptide platform for integrated recognition and reporting: Affinity arrays for protein-ligand interaction analysis

Article abstract of DOI:10.1002/chem.200305391

A molecular platform for protein detection and quantification is reported in which recognition has been integrated with direct monitoring of target-protein binding. The platform is based on a versatile 42-residue helix-loop-helix polypeptide that dimerize

Full text of DOI:10.1002/chem.200305391

L. Baltzer et al.
2375
Chem. Eur. J. 2004, 10, 2375 2385
DOI: 10.1002/chem.200305391
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

FULL PAPER
A Versatile Polypeptide Platform for Integrated Recognition and Reporting:
Affinity Arrays for Protein Ligand Interaction Analysis
[a]
KarinEnander,
Gunnar T. Dolphin,[a] Bo Liedberg,[b] Ingemar Lundstrˆm,[c] and
Lars Baltzer*[a]
Abstract: A molecular platform for
protein detection and quantification is
reported in which recognition has been
integrated withdirect monitoring of
target-protein binding. The platform is
based on a versatile 42-residue helix
loop helix polypeptide that dimerizes
to form four-helix bundles and allows
site-selective modification withrecog-
nition and reporter elements on the
side chains of individually addressable
lysine residues. The well-characterized
interaction between the model target-
protein carbonic anhydrase and its in-
hibitor benzenesulfonamide was used
for a proof-of-concept demonstration.
An affinity array was designed where
benzenesulfonamide derivatives with
aliphatic or oligoglycine spacers and a
fluorescent dansyl reporter group were
introduced into the scaffold. The affini-
ties of the array members for human
carbonic anhydrase II (HCAII) were
determined by titration withthe target
protein and were found to be highly af-
fected by the properties of the spacers
(dissociation constant Kd =0.02 3 mm).
The affinity of HCAII for acetazol-
amide (Kd =4 nm) was determined in a
competition experiment withone of
the benzenesulfonamide array mem-
bers to address the possibility of
screening substance libraries for new
target-protein binders. Also, successful
affinity discrimination between differ-
ent carbonic anhydrase isozymes high-
lighted the possibility of performing
future isoform-expression profiling.
Our platform is predicted to become a
flexible tool for a variety of biosensor
and protein-microarray applications
within biochemistry, diagnostics and
pharmaceutical chemistry.
Keywords: affinity arrays ¥ biosen-
sors ¥ fluorescence ¥ peptide scaf-
folds ¥ protein chips
Introduction
tant aspects of protein-chip design is the choice of capture
9]
molecules. We and others[4 have identified polypeptides as
The development of protein-chip technology, where immobi-
lized capture molecules are used for parallel detection and
quantification of proteins in a high-throughput manner, is
driven by the current need for rapid large-scale analysis of
attractive candidates, as they are robust and easily synthe-
sized by well-established methods and because of the huge
chemical diversity that can be generated in a short time with
modest synthetic cost and effort. The amino acid sequence
can be varied over a wide range to provide access to a varie-
ty of scaffold structures, and highly specific recognition sites
and reporter groups can be introduced postsynthetically by
orthogonal strategies on the solid phase. In our laboratory,
de novo designed 42 amino acid helix loop helix polypepti-
des that undergo pH-controlled site-selective self-functional-
3]
protein function and expression.[1 One of the most impor-
[a] Dr. K. Enander, Dr. G. T. Dolphin, Prof. L. Baltzer
Division of Chemistry
Department of Physics and Measurement Technology
Biology and Chemistry
Linkˆping University, 58183 Linkˆping (Sweden)
Fax : (+46)13-122587
ization withactivated esters ahve been extensively ex-
12]
plored.[10
Function is intimately coupled to structure, as
the fold governs the reactivity of the functionalization sites
through pKa depression of lysine side chains in the hydro-
phobic core and through control of the reactivity of His Lys
pairs. In these scaffolds, a number of lysine residues can be
individually addressed in aqueous solution in an automated
fashion without the need for protecting-group chemistry or
intermediate steps of purification. The fold also provides di-
rectionality and well-defined distances between amino acid
residues. Polypeptide scaffolds designed according to these
principles constitute powerful tools for the construction of,
E-mail: Lars.Baltzer@ifm.liu.se
[b] Prof. B. Liedberg
Division of Sensor Science
Department of Physics and Measurement Technology
Biology and Chemistry
Linkˆping University, 58183 Linkˆping (Sweden)
[c] Prof. I. Lundstrˆm
Division of Applied Physics
Department of Physics and Measurement Technology
Biology and Chemistry
Linkˆping University, 58183 Linkˆping (Sweden)
2376
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200305391
Chem. Eur. J. 2004, 10, 2375 2385

2375 2385
for example, model glycoproteins[13,14] and complex recep-
tors (L. Baltzer et al., unpublished data).
We have previously reported on the design of a functional
biosensor unit based on a helix loop helix polypeptide that
dimerises in solution to form a four-helix bundle.[15] A bio-
sensor provides identification and quantification of a target
substance (analyte) by coupling a highly specific biomolecu-
lar recognition event to changes in physical signals.[16] So-
called affinity biosensors, where the detection signal is a
direct result of analyte binding, frequently employ an optical
transduction mechanism. An attractive approach involves
fluorescence detection, where the events of recognition and
signalling are integrated by attaching a fluorescent reporter
group to the recognition element in a fashion that allows
the reporter to respond to analyte binding.[17]
Microarrays of capture molecules may provide traditional
biosensor signals as well as pattern analysis. We focus on the
design of a flexible molecular platform from which to build
integrated units of recognition and reporting for the analy-
sis, on a chip or in solution, of a variety of protein ligand
systems in an array format. The traditional labelling of the
analyte is avoided since reporting is provided by the capture
agent. In a previous communication, the well-characterized
interaction between the target-protein carbonic anhydrase
(CA) and a benzenesulfonamide ligand was selected for a
proof-of-concept demonstration.[15] The fluorescence of a
peptide modified witha benzenesulfonamide derivative and
a fluorescent probe was monitored as a function of CA con-
centration and the dissociation constant Kd was determined.
Here, we report on the design of a six-membered polypep-
tide affinity array for CA, constructed by varying the spacer
of the benzenesulfonamide derivative. Interaction studies of
the highest-affinity member with different CA isozymes ad-
dress the possibility of protein-expression profiling in a
future diagnostic application, while competition experiments
with the inhibitor acetazolamide show the applicability of
this technology to screening substance libraries for candi-
date drugs.
Figure 1. a) The amino acid sequence of KE2, where lysine residues in
bold represent addressable sites for modification. The amino terminal
was acetylated. Residues that differ from those in the template peptide
LA-42b[13] are underlined. The one letter code for amino acids is used:
A=Ala, D=Asp, E=Glu, F=Phe, G=Gly, H=His, I=Ile, K=Lys,
L=Leu, N=Asn, P=Pro, Q=Gln, R=Arg, V=Val. b) Cartoon of
KE2-D(15)-5 based on the crystal structure of the four-helix-bundle rop
protein (PDB entry code: 1ROP).[45] The rop helices were shortened, the
amino acids were changed to those of KE2 and the dansyl probe 7 and
benzenesulfonamide 5 were attached in positions 15 and 34, respectively
(Sybyl, Tripos Inc.). The polypeptide is folded only when dimerized, but
for reasons of clarity the monomer is shown.
tions 15 and 34 were targeted for postsynthetic modifica-
tions to enhance the probability that the analyte-binding
event would result in interactions between the reporter
group and the analyte because of the proximity and similar
directionality of the incorporated groups in the folded state.
Orthogonal protecting-group chemistry was used for the
functionalization of position 15, while position 34 was acylat-
ed in aqueous solution.
Choice of model system: In the development of a general
biosensor platform, the availability of a well-characterized
model system is crucial for showing proof-of-concept and
for the elucidation of the fundamental characteristics of the
scaffold. The interaction between CA and benzenesulfona-
mide was selected in the development of a KE2-based bio-
sensor molecule.[15] Human carbonic anhydrase II (HCAII),
the most studied of several CA isoforms, is a globular, mon-
omeric enzyme that catalyses the reversible hydration of
carbon dioxide. The enzymatic activity is mediated by a tet-
rahedrally coordinated zinc ion, located in a cavity that is
15-ä deep, where it is ligated by three histidine residues.[21]
Aromatic and heterocyclic sulfonamides are powerful inhibi-
tors of HCAII[22,23] and inhibition is caused by competitive
coordination of the ionized sulfonamide nitrogen atom to
the zinc ion.[24] The interactions between HCAII (and the
bovine equivalent BCAII) and a wide range of benzenesul-
fonamide derivatives have been thoroughly characterized in
Results
Scaffold design: A folded polypeptide was chosen as the
basic building block for the construction of general biosen-
sor molecules, to provide well-defined distances and geome-
tries between the amino acid side chains. The 42-residue
polypeptide SA-42 has been shown previously by NMR
spectroscopy, circular dichroism spectroscopy and analytical
ultracentrifugation to form a helix loop helix motif that di-
merises in solution to form a four-helix bundle.[18,19] The
design of KE2 (Figure 1a) was based on the structure of the
42-residue helix loop helix polypeptide LA-42b,[13] a daugh-
ter sequence of SA-42 that contains five sites for acylation
on the side chains of lysine and ornithine residues. The ver-
satility of LA-42b as a scaffold was demonstrated by the se-
quential and site-selective incorporation of three substitu-
ents into the protein scaffold in aqueous solution without in-
termediate purification.[20] In KE2, 36 out of the 42 amino
acids of LA-42b were conserved. Lysine residues in posi-
31]
the quest for high-affinity inhibitors,[25 and the availability
of structural data for enzyme benzenesulfonamide com-
2377
Chem. Eur. J. 2004, 10, 2375 2385
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

L. Baltzer et al.
FULL PAPER
plexes[30,32] as well as the monovalent character of the inter-
action made it an attractive model system for the evaluation
of the polypeptide-based biosensor platform.
to a larger extent by HCAII binding than the signal from
NBD (data not shown).
5-Dimethylaminonaphthalene-1-sulfonyl chloride (dansyl
chloride, 7, Scheme 1)[33] was chosen for reporting because
when it reacts with amines it forms fluorescent probes char-
acterized by environmentally sensitive fluorescence quan-
tum yields and emission maxima (excitation lmax =335 nm in
methanol).[34,35] Preliminary data withdansyl or 7-nitrobenz-
2-oxa-1,3-diazole (NBD) probes in position 15 of KE2 indi-
cated that the dansyl signal in that position was influenced
Designof anaffinity array : Six derivatives of 4-carboxyben-
zenesulfonamide (1 6, Scheme 1) were synthesized and in-
corporated into the KE2 polypeptide scaffold to form a six-
membered affinity array for HCAII. Benzenesulfonamides
substituted in the para position withalkyl, oligo(ethylene
glycol) or oligopeptide residues are known to bind tighter to
CA than H2NO2SC6H5 does, due to interactions witha hy-
drophobic patch in the enzyme binding cleft.[26 29,31] A gradu-
al increase in the number of methylene groups in the alkyl
chain is accompanied by a decrease in the Kd value,[25,29] but
there is no such correlation when extending the substituent
with more than one ethylene glycol or glycine group.[28] We
therefore selected a series of aliphatic substituents contain-
ing between one and five methylene groups (2 5) in th e
design of the array, expecting steric constraints caused by
the polypeptide scaffold as well as binding-energy contribu-
tions from the spacer methylene groups to influence the af-
finity. Inspection of the crystal structure of HCAII com-
plexed withH 2NO2SC6H4CONH(CH2CH2O)2CH2CH2NH2
(PDB entry 1CNX)[30] suggested that KE2 functionalized
with 5 in position 34 (and possibly some of the ligands with
shorter spacers as well) should be able to reach into the
pocket without distorting the polypeptide structure. A trigly-
cine derivative, 6, was included in the array to allow a com-
parison between the hydrophobic spacer provided by 5 and
a less flexible, more polar one of similar length. The carboxy-
benzenesulfonamide derivative 1 was also included.
Peptide synthesis, functionalization and structural character-
ization: The polypeptide KE2 was synthesized on the solid
phase and a dansyl probe (D) was introduced after selective
deprotection of the amine group on the side chain of Lys15
to form KE2-D(15). The peptide was cleaved from the resin
by 95% trifluoroacetic acid, purified by reversed-phase
HPLC and identified from the MALDI-TOF mass spec-
trometry. The benzenesulfonamide ligands 1 6 were incor-
porated on the side chain of Lys34 by reaction with KE2-
D(15) in aqueous solution at pH 8.0 to form the peptides
KE2-D(15)-X (X=1 6). (KE2-D(15) corresponds to KE2-P
in earlier work[15] and KE2-D(15)-5 corresponds to KE2-
PL.) The purified peptides were identified by MALDI-TOF
mass analysis. Due to the much higher reactivity of Orn34
than of Lys33 in peptides derived from LA-42b,[11] monomo-
dification of KE2 was assumed to have taken place exclu-
sively on the side chain of Lys34. A cartoon of KE2-D(15)-5
is shown in Figure 1b to illustrate the relative positions of
the incorporated ligands.
The secondary structure content of KE2-D(15) and KE2-
D(15)-X (X=1 6) in 100 mm sodium phosphate at pH 7.4
and at the concentration used for fluorescence spectroscopy
studies was determined by circular dichroism spectroscopy
(Table 1). Helical proteins show a characteristic signature
withminima at 208 and 222 nm and all peptides reported
here were found to exhibit a high helical content with mean
residue ellipticities at 222 nm, [V]222, in the range of ꢀ18000
to ꢀ20000 degcm2 dmolꢀ1. The corresponding mean residue
Scheme 1. Structures of the six benzenesulfonamide derivatives used in
the affinity series (1 6), dansyl chloride (7) and the CA inhibitor acet-
azolamide (8).
2378
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 2375 2385

Affinity Array for Carbonic Anhydrase
2375 2385
Table 1. The mean residue ellipticities at 222 nm of KE2, KE2-D(15)
on the Stokes× shifts were observed and the intensity maxi-
mum in the absence of HCAII was 516 518 nm. All pep-
tides except KE2-D(15)-5 displayed a slight blue shift of
ꢁ5 nm when bound to HCAII, a result that indicates a less
polar environment of the probe in the bound peptide than
in the unbound state. The affinities were determined by
plotting the fluorescence intensity at 516 nm versus the total
concentration of HCAII and fitting an equation describing
the dissociation of a bimolecular complex to the experimen-
tal results (Figure 3). The Kd value was confirmed to be
0.02 mm for the interaction between KE2-D(15)-5 and
HCAII, while the affinities of the other peptides were in the
range 0.7 3 mm (Table 2). The affinity of KE2-D(15)-4 was
measured four times (Kd =0.61, 0.77, 0.84 and 0.85 mm) and
that of KE2-D(15)-2 was measured twice (Kd =0.50 and
0.82 mm). From this result, the experimental errors were esti-
mated not to exceed ꢂ20%.
and KE2-D(15)-X (X=1 6).[a]
Peptide
Mean residue ellipticity [V]222
[deg cm2 dmolꢀ1
]
KE2
ꢀ14600
ꢀ19800
ꢀ19600
ꢀ19400
ꢀ19800
ꢀ18600
ꢀ18200
ꢀ20000
KE2-D(15)
KE2-D(15)-1
KE2-D(15)-2
KE2-D(15)-3
KE2-D(15)-4
KE2-D(15)-5
KE2-D(15)-6
[a] Measured at concentrations of 1 mm (or 0.8 mm in the case of KE2-
D(15)-5) in 100 mm sodium phosphate at pH 7.4. The experimental error
was estimated to be ꢂ1000 degcm2 dmolꢀ1
.
ellipticity of KE2 was only ꢀ14600 degcm2 dmolꢀ1 and the
incorporation of substituents thus stabilized the folded struc-
ture, in agreement withwhat was previously observed upon
glycosylation of LA-42b.[13,36] This effect might be due to the
neutralization of the positive charge of the side chain of
Lys34 upon functionalization, but dansyl and benzenesulfo-
namide groups may also stabilise the peptide through inter-
actions within the hydrophobic core. The KE2 scaffold is a
more stable four-helix bundle than LA-42b at low micromo-
lar concentrations, probably due to differences in the amino
acids that build up the peptide core[13] resulting in better hy-
drophobic packing in KE2.
The affinity of H2NO2SC6H5 for HCAII is described by a
dissociation constant, Kd, of 1.5 mm,[25] and para-substituted
Affinity studies: The biosensing ability of KE2-D(15)-5 has
been demonstrated previously and its affinity for HCAII
was 0.02 mm.[15] The basis for biosensing was the fluores-
cence-intensity increase of the dansyl probe upon peptide
binding to HCAII. Here, the affinities of KE2-D(15)-X
(X=1 6) were determined fluorometrically at pH 7.4 and
KE2-D(15) was included as a negative control. Samples of
1 mm peptide (or 0.8 mm in the case of KE2-D(15)-5) were
excited at 335 nm. Upon titration withHCAII (5 n m 90 mm),
the fluorescence intensity increased. The maximum en-
hancements of KE2-D(15)-X (X=1 6) varied between 90%
(KE2-D(15)-1) and 225% (KE2-D(15)-5, Figure 2). No en-
hancement was obtained for KE2-D(15). Only small effects
Figure 3. a) Fluorescence intensity at 516 nm of 1 mm peptide (or 0.8 mm
in the case of KE2-D(15)-5) as a function of HCAII concentration. All
peptides, except the negative control KE2-D(15) (&), display increased
intensities upon addition of HCAII. An equation that describes the disso-
ciation of a bimolecular complex was fitted to the experimental data to
^
^
,
obtain the Kd values: KE2-D(15)-1 ( , Kd =3 mm), KE2-D(15)-2 (
~
*
1 mm), KE2-D(15)-3 ( , 0.7 mm), KE2-D(15)-4 ( , 0.8 mm), KE2-D(15)-5
!
(+, 0.02 mm), KE2-D(15)-6 ( , 3 mm). Average Kd values for KE2-D(15)-
3 (two measurements) and KE2-D(15)-4 (four measurements) are given,
while representative data sets are shown in the figure. Solid lines repre-
sent the best fits to the experimental results. b) Binding curves obtained
by plotting normalized fluorescence data versus the concentration of free
HCAII, [HCAII]free. The concentrations were calculated from the Kd
values.
Figure 2. Fluorescence spectra of 0.8 mm KE2-D(15)-5 as a function of
HCAII concentration. The intensity maximum was at 516 nm.
2379
Chem. Eur. J. 2004, 10, 2375 2385
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

L. Baltzer et al.
FULL PAPER
Table 2. Carbonic anhydrase II affinities for free versus polypeptide-conjugated benzenesulfonamides with ali-
phatic and oligoglycine substituents or spacers. n=number of methylene groups. Cited Kd values for measure-
ments withHCAII and BCAII are included.
ference between HCAI and
HCAII in binding sulfonamides
is ligand-dependent: HCAI
binds dansylamide better that
does HCAII,[37] but HCAII is
superior in binding 2-acetylami-
do-1,3,4-thiadiazole-5-sulfona-
mide (acetazolamide).[38,39]
Sulfonamide
Protein
Kd [mm]
n=1 n=2 n=3
n=4
n=5
H2NO2S(C6H4)CONH(CH2)nCONH-KE2-D(15) HCAII
1
nd
0.03
0.17
0.7[a]
0.8[b]
0.02
[c]
H2NO2S(C6H4)CONH(CH2)nꢀ1CH3
H2NO2S(C6H4)CONH(CH2)nꢀ1CH3
HCAII
BCAII
HCAII
HCAII
0.08
nd
3
0.008 0.0032 0.0018
0.038 0.016 nd
[d]
H2NO2S(C6H4)CO(Gly)3-KE2-D(15)
H2NO2S(C6H4)CO(Gly)4OH[e]
0.36
Acetazolamide
(8,
Scheme 1)[40] is a nanomolar in-
hibitor of CA that has been ex-
haustively studied due to its
[a] Average of two measurements, see text for details. [b] Average of four measurements, see text for details.
[c] From ref. [26]. The affinities were measured in 50 mm tris(hydroxymethyl)aminomethane (Tris)/H2SO4
(pH 8.0). [d] From ref. [29]. The affinities were measured in 20 mm phosphate buffer (pH 7.5) at 378C.
[e] From ref. [31]. The affinities were measured at pH 7.5, the buffer used was not stated.nd=not determined.
benzenesulfonamides resembling 2 5 span the affinity range
of 80 2 nm[26] due to the additional binding energy between
the enzyme binding cleft and the hydrophobic substituents.
In the peptides reported here lower affinities were observed
(1 0.02 mm, Table 2), probably as a result of steric strain
caused by contact between HCAII and the scaffold. The de-
pendence of affinity values on the spacer lengths of peptide-
linked benzenesulfonamide derivatives was very different
from that of similar benzenesulfonamides not coupled to the
peptide scaffold. There was virtually no affinity difference
between the peptides KE2-D(15)-2, KE2-D(15)-3 and KE2-
D(15)-4, but a striking effect of a single methylene group
was observed when comparing the affinities of KE2-D(15)-4
and KE2-D(15)-5. The nanomolar affinity that arises from
cooperativity between zinc-ion coordination and spacer pro-
tein hydrophobic interactions and that is characteristic of
benzenesulfonamide ligands withalipahtic substituents
when not coupled to a scaffold was only observed with five
methylene groups in the spacer. Although this is the most
likely explanation, we do not exclude the possibility that
other parameters might also influence the binding affinities,
for example, interactions between the target protein and the
reporter group, the side chain of Lys34 or the side chains of
other amino acids. The higher the affinity, the less pro-
nounced was the blue shift and the larger was the fluores-
cence difference between free and bound peptide, a result
indicating that the nature of the benzenesulfonamide spacer
influenced the environmental change of the probe.
Benzenesulfonamides witholigoglycine substituents have
been reported previously to bind less tightly to HCAII than
those bearing alkyl groups,[26,28] a tendency that was also ob-
served here. The affinity of KE2-D(15)-6 was two orders of
magnitude lower than that of KE2-D(15)-5 although they
had ligands with similar spacer lengths. As observed for the
aliphatic benzenesulfonamides, the binding affinity of the
peptide-associated glycine derivative was lower than that of
a similar ligand not coupled to the polypeptide scaffold
(Table 2).
Figure 4. Titration of 1 mm KE2-D(15)-5 (or 0.8 mm in the case of HCAII)
~
*
( +, Kd =0.02 mm), BCAII ( , 0.2 mm) and HCAI ( ,
withHCAII
0.01 mm). The solid lines represent the best fit to the experimental data
of an equation that describes the dissociation of a bimolecular complex.
The measurements were performed in duplicate and representative
curves are shown.
early recognition as an antiglaucoma drug. KE2-D(15)-5
and HCAII were mixed in proportions where the degree of
complex formation was at least 98%. When titrating the
sample withacetazolamide, the fluorescence decreased as a
consequence of the increasing concentration of free peptide
due to efficient acetazolamide competition for the HCAII
binding sites (Figure 5). Control experiments where the pep-
tide was titrated withacetazolamide in teh absence of
HCAII revealed that the inhibitor itself caused some minor
fluorescence quenching. Under the assumption that these
two phenomena can be treated additively, the competition
data was compensated for inhibitor quenching by correction
with data from the control experiment. No photobleaching
of the probe was observed during the titration process. The
measurement was performed twice, with2 and 4 mm peptide,
and the affinity of acetazolamide for HCAII was estimated
to be 4 nm from bothmeasurements. Literature values differ
considerably, from 7 70 nm,[27,38,39,41] probably according to
the method and conditions chosen for the affinity analysis.
Apart from confirming that the fluorescence changes of
KE2-D(15)-X upon titration withCA are due to binding of
the benzenesulfonamide-linked peptide to the active site of
the enzyme, this experiment highlights the possibility of
screening substance libraries for new target-protein binders.
The affinity of KE2-D(15)-5 for HCAII (0.02 mm) was
compared to that for the bovine equivalent BCAII (0.2 mm)
as well as to that for the isozyme human carbonic anhydra-
se I (HCAI, 0.01 mm, Figure 4). Our results are in agree-
ment withearlier studies of benzenesulfonamide interac-
[26]
tions withHCAII
and BCAII[29] where tighter binding by
the ligands to HCAII has been observed (Table 2). The dif-
2380
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 2375 2385

Affinity Array for Carbonic Anhydrase
2375 2385
orescent reporter groups are flexibly attached on the side
chains of individually addressable lysine residues. Optimiza-
tion of the capture element for a particular target protein is
aided by the relative ease by which structures, interresidue
distances and relative directionalities of the attached groups
can be varied. This is a general advantage with choosing a
large scaffold, although optimization of all the mentioned
parameters may not always be necessary for the realization
of a functional peptide. Synthetic polypeptides are attractive
capture molecules as they are robust and easy to produce.
For target-protein quantification, an array of capture ele-
ments withdifferent affinities for the analyte is an attractive
alternative to traditional dilution techniques where the
sample concentration must be systematically varied to
matchthe affinities between analytes and capture molecules.
The affinity of a functionalized, folded polypeptide for a
target protein can be varied over several orders of magni-
tude and withrelative ease in comparison withother scaf-
folds. The ligand can be modified by the incorporation of
spacers capable of nonspecific interactions withthe target.
The affinity can also be varied by modification of the scaf-
fold structure, through incorporation of amino acid residues
or groups that enhance or decrease the affinity for the
target due to electrostatic, hydrophobic or steric effects. To
further expand the range of the array, a competing ligand of
appropriate affinity can be added to the sample to decrease
the apparent affinity, with weak-binding array members af-
fected more than tight-binding ones.
In the design of an affinity array for CA, we varied the
para substituents of the benzenesulfonamide ligand to pro-
vide spacers capable of modulating the interactions between
the ligand and the CA binding pocket. We hypothesized
that variation of the spacer structure would provide a range
of binding affinities, but also that the bulky scaffold, which
was not expected to fit into the binding pocket of the
enzyme, would further modulate the interaction due to
steric strain between target protein and the polypeptide;
this would provide a general strategy for the design of affini-
ty arrays. Following this general approach, six array mem-
bers were synthesized and dissociation constants in the
range 10ꢀ8 10ꢀ6 m were obtained. Benzenesulfonamides with
aliphatic para substituents have affinities for CA in the
nanomolar range, because the affinity for the zinc ion in the
active site is enhanced by interactions between the aliphatic
substituent and a hydrophobic patch in the enzyme binding
cleft. The binding strength is increased in a stepwise manner
in proportion to the number of methylene groups.[26,29]
When attached to the polypeptide scaffold, the affinities of
array members bearing one, three and four methylene
groups were found to be similar and approximately 1 mm.
This is most likely a result of steric interference that ob-
scures the manifestation of the intrinsic and gradually in-
creasing affinities of the ligands. However, the affinity dra-
matically increased by two orders of magnitude when five
methylene groups were present in the spacer. The affinity
difference between KE2-D(15)-4 and KE2-D(15)-5 may be
due to a highly cooperative and optimal arrangement of
multiple affinity-enhancing contributors such as the aliphatic
spacer, the fluorescence probe or other residues of the scaf-
Figure 5. Competition for 1 mm HCAII between 2 mm KE2-D(15)-5 and
~
acetazolamide ( ). The observed quenching of the fluorescence of the
dansyl probe withincreased acetazolamide concentration was a result of
the increasing concentration of free KE2-D(15)-5 as HCAII binding sites
became occupied by acetazolamide. However, when the titration was car-
~
ried out in the absence of HCAII ( ), some quenching was still observed
for acetazolamide concentrations above 1 mm. Fluorescence intensities re-
~
corded under titration conditions at [acetazolamide]tot >1 mm ( ) were
compensated by the addition of the difference between the average fluo-
rescence intensities at [acetazolamide]tot ꢃ1 mm and eachrecorded inten-
~
sity for which [acetazolamide]tot >1 mm ( ). The HCAII affinity for acet-
azolamide was estimated to be 4 nm, based on the corrected curve (+).
All fluorescence experiments were performed in a 2-[4-(2-
hydroxyethyl)-1-piperazinyl]ethanesulfonic acid (HEPES)
buffer containing a physiological concentration of NaCl
(0.15m). Most monovalent anions inhibit CA and Clꢀ ions
have a bovine CA binding constant of 0.19m.[42] Although
the concentration of Clꢀ ions in the samples was comparable
to this affinity, the presence of Clꢀ ions only slightly affected
the analysis due to the much tighter binding of the peptides.
When data from measurements with the micromolar-affinity
peptides KE2-D(15)-X (X=1 4, 6) were refitted withan
equation allowing for two competing ligands, one of which
had a Kd value of 0.19m, the apparent affinities were in-
creased by factors of 1.4 1.8. No measurable influence from
the presence of Clꢀ ions on the affinity of the nanomolar-af-
finity peptide KE2-D(15)-5 was expected. Addition of inhib-
itors of appropriate affinities in the sample may therefore
be used for tuning the array by decreasing the apparent af-
finity of weak-binding array members while not affecting
tight-binding ones.
Discussion
The expression of parts of a proteome can be monitored by
bioanalytical arrays of capture molecules suchas antibodies,
aptamers, affibodies, etc.[2] on chips or in microtitre-plate
format. ELISA-like sandwichstrategies or sample labelling,
for example, by radioisotopes, are commonly applied for de-
tection. An appealing alternative is a one-step analysis
where the desired specific recognition is directly integrated
withmonitoring of target-protein binding. We propose a ver-
satile molecular platform based on synthetic, folded helix
loop helix polypeptides where recognition elements and flu-
2381
Chem. Eur. J. 2004, 10, 2375 2385
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

L. Baltzer et al.
FULL PAPER
fold. The affinity profile of benzenesulfonamides with ali-
phatic para substituents changing from stepwise to switch-
like upon incorporation into the scaffold was an unexpected
result, as switch-like affinity behavior is expected from a
rigid rather than from a dynamic protein structure. Substan-
tial flexibility of the protein and/or the polypeptide was, in
fact, demonstrated by specific binding of the sulfonamide
moiety attached to short-spaced ligands that, according to
the crystal structure of HCAII in complex with a benzene-
sulfonamide inhibitor, should not be able to reach deep
enough into the pocket without severe steric strain. The
binding distance would be too long if the structures re-
mained undistorted.[30] A possible explanation for the ob-
served result may be that, if CA is distorted by the scaffold,
the protein structure is no longer complementary to the pre-
ferred conformation of the spacer and the protein ligand in-
teractions that account for the observed nanomolar affinities
of ligands not linked to the scaffold do not arise. With this
interpretation, our findings indicate that minute changes
induce dramatic structural effects in the protein at a critical
spacer length. Nevertheless, the results demonstrate that a
variation in spacer lengthand composition is a general ap-
proachto the design of affinity arrays.
Experimental Section) to the experimental results provides
a direct readout of HCAII concentration, provided that the
sample concentration matches the affinity range of the array
and that start and end plateaus can be determined. The
range of the array affinities should span at least three orders
of magnitude, unless, instead of fitting, the fraction of pep-
tide complexed by HCAII is estimated from the fluores-
cence-intensity change of individual array members upon
complexation. As the peptide concentration is known,
HCAII concentration can be readily determined withthis
strategy provided that calibrated intensities of fully com-
plexed polypeptides are available. Optimal array perform-
ance, where the largest signal differences are obtained be-
tween array members, is obtained when the peptide concen-
tration is less than the lowest Kd value and the protein con-
centration is close to the average of the highest and lowest
Kd values. In Figure 6 the peptide concentration is subopti-
mally 1 mm, which is close to the highest Kd value. Still, for
HCAII concentrations in the micromolar range, concentra-
tion differences of 20 30% are detectable. The reliability of
the analysis is directly related to the number of array mem-
bers.
This discussion is of general applicability to the design of
affinity arrays and illustrates the need for flexible and readi-
ly functionalized sensor platforms. The designed four-helix-
bundle proteins reported here represent versatile scaffolds
in which huge chemical diversity can be generated in a short
time by procedures that are conveniently automated. Array
members withwidely different affinities and reporter prop-
erties are obtained withefforts that are modest in compari-
son withtraditional approaches in organic chemistry and
protein chemistry.
Although five of the array members clustered at low mi-
cromolar affinity, the principle of HCAII-concentration de-
termination in an unknown sample by using the affinity
array can be illustrated (Figure 6). Data from Figure 3, ob-
tained from measurements at a constant total concentration
of HCAII, are presented as normalized fluorescence intensi-
ties of each of the array members plotted versus their affini-
ties. Withenoughdata points, the fitting of an equation de-
scribing the theoretical binding model (Equation (2) in the
Designed proteins are not limited to use as scaffolds for
affinity arrays but are also readily applied to several other
related bioanalytical problems in biochemistry, diagnostics
and pharmaceutical chemistry. Affinity screening of protein
isoforms or mutants was demonstrated withKE2-D(15)- 5 by
discrimination between the three CA isozymes, HCAI,
HCAII and BCAII. If the target protein is an enzyme and
the ligand is a competitive inhibitor, the array also provides
a means of estimating enzyme activity more conveniently
than by traditional assays based on spectrophotometry. Di-
agnostic protein profiling could reveal differences between
the patient and the control with respect to expression levels
of different isoforms of the target protein. As a screening
platform for pharmaceutical purposes, the array could be
challenged with substance libraries to identify new ligands.
The library-screening idea was demonstrated with the acet-
azolamide competition experiments, where the affinity of
the competitive inhibitor was measured. This concept could
also provide a route to new ligands for target proteins from
combinatorial libraries in the case when no tight-binding li-
gands are known.
Figure 6. Normalized fluorescence intensities for different values of
[HCAII]tot plotted versus the Kd values of the KE2-D(15)-X (X=1 6)
array (from Figure 3). Five different [HCAII]tot values were chosen:
!
~
*
*
50 nm ( ), 0.8 mm ( ), 1 mm ( ), 3 mm ( ) and 30 mm (&). Fluorescence
data for KE2-D(15)-5 (Kd =0.02 mm) were collected at [KE2-D(15)-5]tot
=
0.8 mm, but for comparison with the other peptides measured at 1 mm it
was corrected through subtraction of the theoretical difference between
the signal at 0.8 and 1 mm from the original experimental data. The esti-
mated experimental error in the Kd values (ꢂ20%) is indicated for
[HCAII]tot =1 mm. [HCAII]tot =30 mm saturates the array, while 50 nm is
too dilute to induce a response. The optimum concentration is in the low
micromolar range, where data for 0.8 and 1 mm can be resolved. The solid
line is the theoretical binding curve for [HCAII]tot =1 mm when the pla-
teau values are set to 1 and 0.
This work describes fundamental properties of a multi-
functional capture element designed for a variety of biosen-
sor and protein chip applications. Our approach applies to
cases where the protein is known and at least one ligand is
available or can be guessed. Incorporated into the scaffold,
the known ligand may be the starting point for screening
2382
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 2375 2385

Affinity Array for Carbonic Anhydrase
2375 2385
(1:3) with4% acetic acid, Rf =0.6). The solution was acidified to pH 3
and reduced in vacuo. The intermediate carboxylic acid product,
substance libraries for better ligands. An affinity array can
be designed by varying specific or nonspecific ligand protein
interactions that arise from variation of the ligand, the
spacer or the scaffold. Depending on the interaction studied
and on the application, issues of fluorophore and ligand
array optimization, peptide immobilization or nonspecific
binding of other proteins present in a blood sample or lysate
may call for solutions. Finding these solutions will be facili-
tated by the flexibility of the polypeptide scaffold in terms
of amino acid variation, as well as options for postsynthetic
modifications.
H2NO2S(C6H4)CONH(CH2)nCOOH
(n=1,
3,
4,
5)
or
H2NO2S(C6H4)CO(Gly)3, was washed in cold water and dried. Yields
1
were 62 79% and the purity was confirmed by H NMR spectroscopy. To
prepare the NHS ester, the intermediate product (1 equiv) was added to
0.1m NHS (1.1 equiv) in dioxane:DMF (1:1) at 08C. DCC (1 equiv) was
added to the solution and the reaction mixture was stirred overnight at
58C. DCU was removed by filtration. The solvent was reduced and a
white solid (2 6) was precipitated from 2-propanol and hexane. The
purity was confirmed by 1H NMR spectroscopy and the yields were 36
66%. A typical synthetic procedure is provided below.
H2NO2S(C6H4)CONH(CH2)5CO2(NC4H4O2) (5): 6-Aminohexanoic acid
(88 mg, 0.67 mmol) in 50 mm sodium borate (2 mL) was added to 1
(200 mg, 0.67 mmol) in acetone (3 mL) at pH 8.5. The reaction mixture
was stirred at room temperature for two hours while keeping the pH
value between 7.0 and 7.5, after which the solution was acidified to pH 3
and concentrated. The intermediate product H2NO2S(C6H4)-
CONH(CH2)5COOH was washed in cold water and vacuum dried to give
a white solid (145 mg, 69%); 1H NMR (300 MHz, D2O, 25 8C): d=8.01
(d, 2H), 7.91 (d, 2H), 3.37 (t, 2H), 2.17 (t, 2H), 1.58 (m, 4H), 1.36 (m,
2H) ppm. H2NO2S(C6H4)CONH(CH2)5COOH (100 mg, 320 mmol) and
NHS (41 mg, 352 mmol) were dissolved in a mixture of dioxane (1 mL)
and DMF (1 mL) and the mixture was cooled in an ice bath. DCC
(65 mg, 320 mmol) was added to the solution and the mixture was stirred
overnight at 58C. DCU was removed by filtration. The solvent was re-
duced to an oil and a white solid was precipitated from 2-propanol and
hexane. Recrystallization from hot 2-propanol afforded the desired prod-
Conclusion
This work proposes the use of designed, functionalized poly-
peptides as efficient capture elements for protein detection
and quantification in an analytical array format. The relative
ease by which individually addressable lysine side chains
may be functionalized facilitates optimization of the plat-
form for the interaction system of interest, while integration
of the processes of recognition and reporting allows one-
step analyses of protein abundance, activity and affinity. For
the model system of carbonic anhydrase and benzenesulfo-
namide, an affinity array was designed where, at a critical
benzenesulfonamide spacer length, subtle spacer differences
induced a profound effect on the affinity. Discrimination be-
tween different carbonic anhydrase isoforms and a concept
for finding new ligands by competition studies was also
demonstrated. These experiments are milestones on the way
towards the design of a protein chip with immobilized cap-
ture polypeptides, where issues within biochemistry, diagnos-
tics or pharmaceutical chemistry can be addressed.
1
uct (5, 47 mg, 36%); H NMR (300 MHz, [D6]acetone, 258C): d=8.02 (d,
2H), 7.95 (d, 2H), 7.87 (s, 1H), 6.66 (s, 2H), 3.44 (q, 2H), 2.88 (s, 4H),
2.66 (t, 2H), 1.78 (m, 2H), 1.69 (m, 2H), 1.54 (m, 2H) ppm.
Peptide synthesis, purification and modification: The polypeptide KE2
was synthesized on a Pioneer automated peptide synthesiser (Applied Bi-
osystems) by using standard 9-fluorenylmethoxycarbonyl (Fmoc) chemis-
try. The synthesis was performed on a 0.2-mmol scale and an Fmoc-Gly
polyethylene glycol polystyrene polymer (Applied Biosystems) with a
substitution level of 0.18 mmolgꢀ1 was used. This resin was purchased
with the first amino acid attached and the free acid was formed at the C
teminus upon cleavage in trifluoroacetic acid (TFA). The side chains of
the amino acids (Calbiochem Novabiochem AG) were protected by
base-stable groups: tert-butyl ester (Asp, Glu), tert-butoxymethyl (Lys),
trityl (His, Asn, Gln) and 2,2,4,6,7-pentamethyldihydrobenzofuran-5-sul-
fonyl (Arg). To allow site-selective incorporation of a fluorescent probe
at position 15, the side chain of Lys15 was protected by an allyloxycar-
bonyl (Applied Biosystems) group that can be selectively removed by
treatment with tetrakis(triphenylphosphine)palladium(0). The Fmoc pro-
tecting groups were removed from the amino termini by treatment with
20% piperidine in DMF. A fourfold excess of amino acid was used in
eachcoupling and amino acids were activated witha mixture of O-(7-
benzotriazole-1-yl)-1,1,3,3-tetramethyluronium tetrafluoroborate (TBTU,
0.5m in DMF; Alexis Biochemicals) and diisopropylethylamine (DIPEA,
1m in DMF). A standard amino acid coupling time of 60 min was used,
except in the cases of Gln and His (90 min) and Asn and Arg (120 min).
The N terminus of the peptide was capped with 0.3m acetic anhydride in
DMF. After the completed synthesis, the resin was rinsed with dichloro-
methane (DCM) and dried in vacuo.
Experimental Section
The origin of chemicals obtained from standard chemical suppliers is not
stated.
Synthesis of benzenesulfonamide derivatives: Active esters of inhibitors
were prepared following procedures described previously for similar
compounds,[28] withminor modifications. 4-Carboxybenzenesulfonamide
was used as the starting material for all inhibitors. Yields were not opti-
mized.
H2NO2S(C6H4)CO2(NC4H4O2) (1): 4-Carboxybenzenesulfonamide (2.0 g,
10 mmol) and N-hydroxysuccinimide (NHS, 1.14 g, 10 mmol) were dis-
solved in a mixture of dioxane (20 mL) and N,N-dimethylformamide
(DMF, 8 mL) and the mixture was cooled in an ice bath. Dicyclohexylcar-
bodiimide (DCC, 2.0 g, 10 mmol) was added to the solution and the reac-
tion mixture was stirred overnight at 58C. N,N’-dicyclohexylurea (DCU)
was removed by filtration. The solvent was reduced to an oil and crystal-
lization from 2-propanol and hexane afforded a white solid (2.1 g, 73%);
1H NMR (300 MHz, D2O, 25 8C): d=8.35 (d, 2H), 8.11 (d, 2H), 3.03 (s,
4H) ppm.
Before cleaving the peptide from the resin, Lys15 was deprotected over
3 hat room temperature by using [Pd(PPh 3)4] (3 equiv) in a mixture of
trichloromethane, acetic acid and N-methylmorpholine (17:2:1 v/v;
12 mL per g of polymer). The resin was washed sequentially with 20 mm
diethyldithiocarbamic acid in DMF, 30 mm DIPEA in DMF, DMF and
DCM, and then desiccated. For probe coupling, the resin was mixed with
dansyl chloride (2 equiv) and DIPEA (7 equiv) in DMF (6 mL per g of
polymer) and the mixture was left to stand with occasional swirling for
4 hat room temperature. The resin was washed withDMF and DCM and
then desiccated. The dansylated peptide was cleaved from the resin by
treatment witha mixture of TFA, triisopropylsilane, and water
(95:2.5:2.5 v/v; 15 mL per g of polymer) over 2 hat room temperature.
After filtration, the TFA was evaporated and the peptide was precipitat-
ed by addition of cold diethyl ether, centrifuged, washed in diethyl ether
and lyophilized. The crude product was purified by reversed-phase
General procedure for the synthesis of H2NO2S(C6H4)CONH(CH2)n-
CO2(NC4H4O2) (2 5: n=1, 3, 4, 5) and H2NO2S(C6H4)CO(Gly)3-
(NC4H4O2) (6): 0.3m Nucleophile (Gly, 4-aminobutyric acid, 5-aminopen-
tanoic acid, 6-aminohexanoic acid or the tripeptide glycyl glycyl glycine,
respectively, 1 equiv) in 50 mm sodium borate was added to 0.2m 1 in ace-
tone at pH 8.5. The reaction mixture was stirred at room temperature for
at least two hours while the pH value was kept between 7.0 and 7.5 with
0.1m NaOH. The reaction can be monitored by thin-layer chromatogra-
phy by observing the disappearance of unreacted 1 (toluene:ethyl acetate
2383
Chem. Eur. J. 2004, 10, 2375 2385
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

L. Baltzer et al.
FULL PAPER
HPLC on a semipreparative HICHROM C-8 column, eluted isocratically
Fitting was done withIGOR Pro 4.03 software (WaveMetrics Inc.). For
estimation of the influence of Clꢀ ions on the measured affinities, Equa-
tion (2) was used withthe expression for [CA] given in Equation (4),
where [Clꢀ]tot is the total concentration of chloride ions, [CAP] is the
concentration of CA complexed by the peptide and KCl is the affinity
constant of chloride-ion binding to CA. [CAP] was estimated from the
experimental data.
with40% 2-propanol and 0.1% TFA in water at
a flow rate of
10 mLminꢀ1. The purified peptide was identified by MALDI-TOF mass
spectrometry.
Benzenesulfonamides 1 6 (1.3 equiv) were added to 1 mm peptide solu-
tions (prepared from lyophilized peptide under the assumption that it
contains 25% water) in 50 mm Tris buffer (pH 8.0) to create the six-mem-
bered peptide array. The incorporation reaction was monitored by
MALDI-TOF mass spectrometry. After one day at room temperature
the reaction mixtures of peptides modified with 1 5 were purified by re-
½CAŠtotꢀ½CAPŠꢀ½ClꢀŠtotꢀKCl
½CAŠ ¼
2
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ð4Þ
ꢀ
ꢁ
versed-phase HPLC on
a semipreparative HICHROM C-8 column,
2 þ KClð½CAŠtotꢀ½CAPŠÞ
½CAŠtotꢀ½CAPŠꢀ½ClꢀŠtotꢀKCl
eluted isocratically with40% 2-propanol and 0.1% TFA in water at a
flow rate of 10 mLminꢀ1. The peptide modified with 6 was purified in the
same way, except that a gradient of 30!70% acetonitrile and 0.1% TFA
in water was applied over 60 min. Monomodification yields were 44
63%, while 10% dimodification (Lys33, Lys34) was observed.
þ
2
Competitionexperiments with acetazolamide : Mixtures of 1 mm HCAII
and 2 mm E2-D(15)-5 or 1 mm HCAII and 4 mm KE2-D(15)-5 in HEPES-
buffered saline (10 mm HEPES, 150 mm NaCl, pH 7.4; Biacore AB) were
titrated withsolutions of 10 and 100 mm acetazolamide. The excitation
wavelength was 335 nm and the emission was monitored in the range
450 650 nm. Excitation and emission slits were 5 nm. Eachspectrum rep-
resents an average of two scans. Control experiments were performed
where 2 or 4 mm peptide was titrated withacetazolamide in the absence
of protein. The dissociation constant Ki of the interaction between KE2-
D(15)-5 and acetazolamide was estimated from the total acetazolamide
concentration [A]tot at which the fluorescence was intermediate between
its maximum and minimum values, that is, where the concentration of
peptide-bound HCAII equals the concentration of acetazolamide-bound
HCAII. At this point, provided the concentration of free protein is ne-
glected, Equation (5) is applicable, where Kd is the dissociation constant
of the peptide protein complex and [P]tot is the total peptide concentra-
tion.
Concentration determination of proteins and peptides: HCAI, HCAII
and BCAII were obtained from Sigma and were used without further pu-
rification. The same protein batches were used for all experiments. Pro-
tein concentrations were determined from the absorbance at 280 nm by
using extinction coefficients of 46800 (HCAI), 54800 (HCAII) and
55100mꢀ1 cmꢀ1 (BCAII).[43] Concentrations of peptide stock solutions di-
luted 1:10 in methanol were determined from the absorbance at 335 nm
by using an extinction coefficient of 4200mꢀ1 cmꢀ1 (dansyl)[35] and were
subsequently confirmed by amino acid analysis (Aminosyraanalyscentra-
len, Uppsala, Sweden).
Structure determination with circular dichroism spectroscopy: Circular
dichroism spectra were recorded on a CD6 spectrodichograph (Jobin-
Yvon Instruments SA) by using a 10-mm cuvette in the interval 260
203 nm at room temperature. Spectra of 1 mm (or 0.8 mm in the case of
KE2-D(15)-5) peptide in 100 mm sodium phosphate at pH 7.4 were ob-
tained by collecting data at 0.5 nm intervals withan integration time of
1 s. Each spectrum was collected as an average of three scans and the
data were corrected by subtraction of a spectrum of a blank sample con-
taining only buffer. The helical content was determined as the mean resi-
½HCAIIŠ
2
tot
½AŠtotꢀ
½PŠtotꢀ
Ki ¼ Kd
ð5Þ
½HCAIIŠ
tot
2
due ellipticity at 222 nm, [V]222, and was calculated from Equation (1),
Acknowledgments
obs
where V is the observed ellipticity at 222 nm (deg), mrw is the mean
222
residue weight (gmolꢀ1), c is the peptide concentration (gmLꢀ1) and l is
K.E. is enrolled in the graduate school Forum Scientium and in the re-
searchprogram Biomimetic Materials Science, bothfinanced by teh
SwedishFoundation for Strategic Research(SSF). Financial support
from the Swedish Research Council (VR) is also gratefully acknowl-
edged.
the optical path length of the cell (cm).
Vo22b2smrw
10lc
ð1Þ
½VŠ
¼
222
Affinity determination with fluorescence spectroscopy: Fluorescence
spectra were recorded on a F-4500 spectrofluorophotometer (Hitachi
High-Technologies Corp.) by using a 10-mm quartz cuvette at 238C. The
excitation wavelength was 335 nm and the emission was monitored in the
range 450 650 nm. Excitation and emission slits were 5 nm. Eachspec-
trum represents an average of two scans. For affinity determinations,
samples of 1 mm (or 0.8 mm in the case of KE2-D(15)-5) peptide in
HEPES-buffered saline (10 mm HEPES, 150 mm NaCl, pH 7.4; Biacore
AB) were titrated withsolutions of 1, 10 and 100 mm CA (HCAI, HCAII
or BCAII) diluted from a stock solution of 500 mm CA in HEPES-buf-
fered saline. The fluorescence intensity at 516 nm was monitored as a
function of the total protein concentration and the dissociation constant
Kd was determined by fitting Equation (2) to the experimental results
under the assumption of a 1:1 binding model.
[1] T. Kodadek, Chem. Biol. 2001, 8, 105.
[2] M. F. Templin, D. Stoll, M. Schrenk, P. C. Traub, C. F. Vohringer,
T. O. Joos, TrendsBiotechnol. 2002, 20, 160.
[3] H. Zhu, M. Snyder, Curr. Opin. Chem. Biol. 2003, 7, 55.
[4] L. Bracci, L. Lozzi, B. Lelli, A. Pini, P. Neri, Biochemistry 2001, 40,
6611.
[5] F. Toepert, J. R. Pires, C. Landgraf, H. Oschkinat, J. Schneider-Mer-
gener, Angew. Chem. 2001, 113, 922; Angew. Chem. Int. Ed. 2001,
40, 897.
[6] B. T. Houseman, J. H. Huh, S. J. Kron, M. Mrksich, Nat. Biotechnol.
2002, 20, 270.
[7] M. Takahashi, K. Nokihara, H. Mihara, Chem. Biol. 2003, 10, 53.
[8] J. L. Naffin, Y. Han, H. J. Olivos, M. M. Reddy, T. Sun, T. Kodadek,
Chem. Biol. 2003, 10, 251.
[9] A. Q. Emili, G. Cagney, Nat. Biotechnol. 2000, 18, 393.
[10] K. Broo, L. Brive, A. C. Lundh, P. Ahlberg, L. Baltzer, J. Am.
Chem. Soc. 1996, 118, 8172.
Fbound½CAŠ þ FfreeKd
½44Š
ð2Þ
Fobs
¼
½CAŠ þ Kd
In Equation (2), Fobs is the observed fluorescence intensity, Fbound is the
fluorescence of the peptide bound to CA, Ffree is the fluorescence of the
free peptide and [CA] is the concentration of free CA. [CA] can be de-
rived from Equation (3), where [P]tot is the total concentration of peptide
and [CA]tot is the total concentration of CA.
[11] L. K. Andersson, G. T. Dolphin, L. Baltzer, ChemBioChem 2002, 3,
741.
[12] L. K. Andersson, M. Caspersson, L. Baltzer, Chem. Eur. J. 2002, 8,
3687.
[13] L. K. Andersson, G. Stenhagen, L. Baltzer, J. Org. Chem. 1998, 63,
1366.
[14] S. Vijayalekshmi, S. K. George, L. K. Andersson, J. Kihlberg, L.
Baltzer, Org. Biomol. Chem. 2003, 1, 2455.
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ꢀ
ꢁ
½PŠtot þ Kdꢀ½CAŠ
½PŠtot þ Kdꢀ½CAŠ
tot
tot
½CAŠ ¼ ꢀ
þ
2 þ Kd½CAŠ
tot
2
2
[15] K. Enander, G. T. Dolphin, L. K. Andersson, B. Liedberg, I. Lund-
strˆm, L. Baltzer, J. Org. Chem. 2002, 67, 3120.
ð3Þ
2384
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2004, 10, 2375 2385

Affinity Array for Carbonic Anhydrase
2375 2385
[16] D. R. Thevenot, K. Toth, R. A. Durst, G. S. Wilson, Pure Appl.
Chem. 1999, 71, 2333.
[17] H. W. Hellinga, J. S. Marvin, TrendsBiotechnol. 1998, 16, 183.
[18] S. Olofsson, G. Johansson, L. Baltzer, J. Chem. Soc. Perkin Trans. 2
1995, 2047.
[34] J. R. Lakowicz, Principles of Fluorescence Spectroscopy, Kluwer
Academic/Plenum Publishers, New York, 1999.
[35] R. P. Haugland, Molecular Probes: Handbook of Fluorescent Probes
and Research Chemicals, to be found under http://www.probes.com/
handbook, 2003.
[19] S. Olofsson, L. Baltzer, Folding Des. 1996, 1, 347.
[20] L. K. Andersson, PhD thesis, Gˆteborg University, Sweden, 2001.
[21] A. Liljas, K. HÂkansson, B.-H. Jonsson, Y. F. Xue, Eur. J. Biochem.
1994, 219, 1.
[36] L. K. Andersson, G. T. Dolphin, J. Kihlberg, L. Baltzer, J. Chem.
Soc. Perkin Trans. 2 2000, 459.
[37] C. Engstrand, B. H. Jonsson, S. Lindskog, Eur. J. Biochem. 1995,
229, 696.
[22] T. Mann, D. Keilin, Nature 1940, 146, 164.
[23] H. A. Krebs, Biochem. J. 1948, 43, 525.
[38] O. Arslan, O. I. Kufrevioglu, B. Nalbantoglu, Bioorg. Med. Chem.
1997, 5, 515.
[24] S. Lindskog, P. J. Wistrand in Design of Enzyme Inhibitors as Drugs
(Eds.: M. Sandler, H. J. Smith), Oxford University Press, Oxford,
1988.
[25] P. W. Taylor, R. W. King, A. S. V. Burgen, Biochemistry 1970, 9,
2638.
[39] C. T. Supuran, A. Scozzafava, J. Enzyme Inhib. 2000, 15, 597.
[40] J. O. J. Roblin, J. W. Clapp, J. Am. Chem. Soc. 1950, 72, 4890.
[41] T. H. Maren in The Carbonic Anhydrases: New Horizons (Eds.:
W. R. Chegwidden, N. D. Carter, Y. H. Edwards), Birkh‰user
Verlag, Basel, 2000.
[26] R. W. King, A. S. V. Burgen, Proc. R. Soc. London Ser. B 1976, 193,
107.
[27] A. Jain, G. M. Whitesides, R. S. Alexander, D. W. Christianson, J.
Med. Chem. 1994, 37, 2100.
[42] Y. Pocker, J. T. Stone, Biochemistry 1968, 7, 2936.
[43] S. Lindskog, L. E. Henderson, K. K. Kannan, A. Liljas, P. O. Nyman,
B. Strandberg in The Enzymes, Vol 5 (Ed.: P. D. Boyer), Academic
Press, New York, 1971.
[28] A. Jain, S. G. Huang, G. M. Whitesides, J. Am. Chem. Soc. 1994, 116,
5057.
[29] J. M. Gao, S. Qiao, G. M. Whitesides, J. Med. Chem. 1995, 38, 2292.
[30] P. A. Boriack, D. W. Christianson, J. Kingerywood, G. M. White-
sides, J. Med. Chem. 1995, 38, 2286.
[44] A. Fersht, Structure and Mechanism in Protein Science: A Guide to
Enzyme Catalysis and Protein Folding, W. H. Freeman, New York,
1999.
[45] D. W. Banner, M. Kokkinidis, D. Tsernoglou, J. Mol. Biol. 1987, 196,
657.
[31] G. B. Sigal, G. M. Whitesides, Bioorg. Med. Chem. Lett. 1996, 6, 559.
[32] A. E. Eriksson, P. M. Kylsten, T. A. Jones, A. Liljas, Proteins 1988, 4,
283.
Received: July 26, 2003
Revised: January 5, 2004 [F5391]
[33] G. Weber, Biochem. J. 1952, 51, 155.
Published online: April 15, 2004
2385
Chem. Eur. J. 2004, 10, 2375 2385
www.chemeurj.org
¹ 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim