High-affinity adaptors for switchable recognition of histidine-tagged proteins

Article abstract of DOI:10.1021/ja050690c

We aspired to create chemical recognition units, which bind oligohistidine tags with high affinity and stability, as tools for selectively attaching spectroscopic probes and other functional elements to recombinant proteins. Several supramolecular entities containing 2-4 nitrilotriacetic acid (NTA) moieties were synthesized, which additionally contained an amino group, to which fluorescein was coupled as a sensitive reporter probe. These multivalent chelator heads (MCH) (termed bis-, tris-, and tetrakis-NTA) were characterized with respect to their interaction with hexahistidine (H6)- and decahistidine (H10)-tagged targets. Substantially increased binding stability with increasing number of NTA moieties was observed by analytical size exclusion chromatography. The binding enthalpies as determined by isothermal titration calorimetry increased nearly additively with the number of possible coordinative bonds between chelator heads and tags. Yet, a substantial excess of histidines in the oligohistidine tag was required for obtaining fully additive binding enthalpies. Dissociation kinetics of MCH/oligohistidine complexes measured by fluorescence dequenching showed an increase in stability by 4 orders of magnitude compared to that of mono-NTA, and subnanomolar affinity was reached for tris-NTA. The gain in free energy with increasing multivalency was accompanied by an increasing loss of entropy, which was ascribed to the high flexibility of the binding partners. Numerous applications of these MCHs for noncovalent, high affinity, yet reversible tethering of spectroscopic probes and other functional elements to the recombinant proteins can be envisioned.

Full text of DOI:10.1021/ja050690c

Published on Web 06/28/2005
High-Affinity Adaptors for Switchable Recognition of
Histidine-Tagged Proteins
Suman Lata,† Annett Reichel,† Roland Brock,‡ Robert Tampe´,† and Jacob Piehler*,†
Contribution from the Institute of Biochemistry, Johann Wolfgang Goethe-UniVersity, Frankfurt/
Main, Germany, and Institute for Cell Biology, Eberhard Karls-UniVersity, Tu¨bingen, Germany
Received February 2, 2005; E-mail: j.piehler@em.uni-frankfurt.de
Abstract: We aspired to create chemical recognition units, which bind oligohistidine tags with high affinity
and stability, as tools for selectively attaching spectroscopic probes and other functional elements to
recombinant proteins. Several supramolecular entities containing 2-4 nitrilotriacetic acid (NTA) moieties
were synthesized, which additionally contained an amino group, to which fluorescein was coupled as a
sensitive reporter probe. These multivalent chelator heads (MCH) (termed bis-, tris-, and tetrakis-NTA)
were characterized with respect to their interaction with hexahistidine (H6)- and decahistidine (H10)-tagged
targets. Substantially increased binding stability with increasing number of NTA moieties was observed by
analytical size exclusion chromatography. The binding enthalpies as determined by isothermal titration
calorimetry increased nearly additively with the number of possible coordinative bonds between chelator
heads and tags. Yet, a substantial excess of histidines in the oligohistidine tag was required for obtaining
fully additive binding enthalpies. Dissociation kinetics of MCH/oligohistidine complexes measured by
fluorescence dequenching showed an increase in stability by 4 orders of magnitude compared to that of
mono-NTA, and subnanomolar affinity was reached for tris-NTA. The gain in free energy with increasing
multivalency was accompanied by an increasing loss of entropy, which was ascribed to the high flexibility
of the binding partners. Numerous applications of these MCHs for noncovalent, high affinity, yet reversible
tethering of spectroscopic probes and other functional elements to the recombinant proteins can be
envisioned.
Introduction
ity of labeling relies upon the recognition fidelity in these
adaptors. Next to selectivity and site-specificity, ideal adaptors
In the current era of proteome research, generic tools for
labeling, handling, modifying, and organizing proteins are
increasingly demanded. During recent years, numerous meth-
odologies for selective, site-specific conjugation of proteins with
reporter molecules such as organic fluorophores,1 spin labels,2
nanoparticles,3 or solid phases4 have been reported. While
several techniques for site-specific covalent coupling with a
diverse range of functional elements are available,5-7 non-
covalent tethering through biological adaptor modules has most
frequently served as an engineering principle for a selective and
specific attachment to proteins.8-14 The selectivity and specific-
for noncovalent modification of proteins should also comply
with the following requirements: (i) stoichiometrically defined
and stable, yet reversible or even switchable binding with the
protein under physiological conditions, (ii) site-specific and
stoichiometric covalent coupling of the spectroscopic reporters
or any other functional element by organic synthetic means,
and (iii) small size for minimum effect on biochemical and
physicochemical properties. Thus, ideal adaptors on one hand
should be well compatible with organic synthesis yet soluble
and stable under aqueous conditions.
Biological recognition units such as antibodies have reached
high affinity and unmatched specificity toward their targets by
an optimal interplay of many weak noncovalent forces such as
hydrophobic interactions, hydrogen bonding, electrostatics, π-π
stacking, cation-π interaction, and so forth. However, proteins
are large and chemically multifunctional macromolecules, which
are incompatible with specific (organic) conjugation chemistry.
Therefore, stoichiometrically and structurally well-defined at-
tachment of spectroscopic probes is hardly possible. Further-
more, high-affinity binding of protein-based recognition units
is often only reversible under drastic denaturing conditions
† Johann Wolfgang Goethe-University.
‡ Eberhard Karls-University.
(1) Holmes, K. L.; Lantz, L. M. Methods Cell Biol. 2001, 63, 185-204.
(2) Hubbell, W. L.; Gross, A.; Langen, R.; Lietzow, M. A. Curr. Opin. Struct.
Biol. 1998, 8, 649-656.
(3) Salata, O. J. Nanobiotechnol. 2004, 2, 3.
(4) Cooper, M. A. Anal. Bioanal. Chem. 2003, 377, 834-842.
(5) Yin, J.; Liu, F.; Li, X.; Walsh, C. T. J. Am. Chem. Soc. 2004, 126, 7754-
7755.
(6) George, N.; Pick, H.; Vogel, H.; Johnsson, N.; Johnsson, K. J. Am. Chem.
Soc. 2004, 126, 8896-8897.
(7) Griffin, B. A.; Adams, S. R.; Tsien, R. Y. Science 1998, 281, 269-272.
(8) Farinas, J.; Verkman, A. S. J. Biol. Chem. 1999, 274, 7603-7606.
(9) Marks, K. M.; Braun, P. D.; Nolan, G. P. Proc. Natl. Acad. Sci. U.S.A.
2004, 101, 9982-9987.
(10) Wu, M. M.; Llopis, J.; Adams, S.; McCaffery, J. M.; Kulomaa, M. S.;
Machen, T. E.; Moore, H. P.; Tsien, R. Y. Chem. Biol. 2000, 7, 197-209.
(11) Chen, I.; Ting, A. Y. Curr. Opin. Biotechnol. 2005, 16, 35-40.
(12) Chen, I.; Howarth, M.; Lin, W.; Ting, A. Y. Nat. Methods 2005, 2, 99-
104.
(13) Miller, L. W.; Cornish, V. W. Curr. Opin. Chem. Biol. 2005, 9, 56-61.
(14) Miller, L. W.; Sable, J.; Goelet, P.; Sheetz, M. P.; Cornish, V. W. Angew.
Chem., Int. Ed. 2004, 43, 1672-1675.
9
10.1021/ja050690c CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 10205-10215
10205

A R T I C L E S
Lata et al.
because several different types of weak forces have to be
interfered with simultaneously. Fast switchability of protein
interaction (e.g., in signal transduction) therefore relies on
complex conformational changes, which typically involve
enzymatic processes or binding of ligands or metal ions. Without
induced conformational changes, the fastest rate at which the
switching may occur is the inherent dissociation rate constant
of a given interaction, thereby making the stable yet switchable
binding elements paradoxical.
NTA has been observed.27,28,39,40 Apparently, stable interaction
at interfaces is possible by multipoint attachment40-42 (i.e.,
multivalent interaction) since the hexahistidine tag provides
binding sites for three NTA moieties and larger oligohistidine
tags even more. Indeed, cyanine dyes conjugated with two NTA
moieties were shown to bind hexahistidine tags in solution with
increased affinity.36
Here, we aimed to design high-affinity adaptors for oligo-
histidine-tagged proteins by incorporating multiple NTA moi-
eties into single supramolecular entities, which can be site-
specifically coupled with spectroscopic or biochemical probes
using organic synthetic chemistry. We used dendritic and cyclic
scaffolds for synthesizing such multivalent chelator headgroups
(MCH) with two, three, and four NTA moieties. By conjugating
these MCHs with a fluorescent dye, we incorporated a reporter,
which was useful for concentration determination and for
studying the kinetics of the interaction with hexa- and deca-
histidine tags. Furthermore, the thermodynamics of these
interactions were assessed by isothermal titration calorimetry
(ITC). The tremendous increase in complex stabilities achieved
by these MCHs enabled stable and stoichiometric conjugation
of proteins with affinities reaching subnanomolar levels.
While designing chemical recognition elements with speci-
ficities comparable to their biological counterparts is currently
beyond scientific capabilities, multivalent interactions have been
successfully exploited for achieving high stabilities of supra-
molecular ligand/receptor complexes.15-23 Another key advan-
tage of multivalent interaction is the possibility to efficiently
diminish its stability to that of the individual interaction by
adding a monovalent competitor,15 enabling for fast dissociating
multivalent complexes under mild conditions. For designing
multivalent chemical receptors for recognizing short peptide
sequences, coordinative interaction is ideally suited: it is
optimally tight compared to other weak interactions, independent
of ionic strength, and relatively specific as only the typically
rarely occurring amino acids histidine and cysteine exhibit
substantial affinity toward transition metal ions. Transition
metals chelated by nitrilotriacetic acid (NTA) or other chelators
have been successfully applied for purification24,25 and detec-
tion25,26 of oligohistidine-tagged proteins, as well as for im-
mobilization on surfaces27-31 and for tethering to lipid
membranes.32-34 Fast dissociation of this interaction under mild
conditions is readily achieved by adding competitors such as
imidazole. Furthermore, the interaction may be switched by
adding or removing the metal ion. While individual Me:NTA-
oligohistidine complexes (which are typically formed by two
coordination bonds) are of very low affinity (KD ≈ 10 µM) and
stability,35-38 stable binding to surfaces densely covered with
Experimental Section
Synthesis of the MCH. The synthesis of the multivalent chelator
heads is outlined in Scheme 1; the details are given in Supporting
Information.
t-Butyl ester-protected NTA synthons (3 and 12) were derived from
commercially available R-amino acid derivatives. The amino-function-
alized synthon 3 was obtained from an appropriately protected L-lysine
derivative as described earlier,35 and the carboxy-functionalized synthon
12 was accordingly obtained from a protected L-glutamic acid. These
synthons were then grafted onto structurally different scaffolds by amide
bond formation. The other functional groups (amine and carboxyl group)
while interfering were kept protected orthogonally to tert-butyl esters
by conversion into benzyl carbamate and benzyl ester, respectively.
Synthon 3 was doubled by coupling to the carboxyl groups of an
L-glutamic acid scaffold (4), resulting in amino-functionalized 6 and
the second-generation amino-functionalized 8. Synthon 12 was grafted
in triplicate onto a tetraaza cyclam (9) under stoichiometric control to
yield 13. The chelator head precursors (3, 6, 8, and 13) were further
modified to present a primary amine functionality followed by
deprotection of the NTA heads (16, 17, 18, and 19).
Adaptor/Reporter Conjugates. As a sensitive probe for monitoring
interactions, fluorescein was coupled to the primary amino group of
the MCH 16, 17, 18, and 19 by using a carboxyfluorescein NHS ester
(Fluka). The products were purified by preparative thin-layer chroma-
tography and by anion exchange chromatography (HiTrap Q, Amersham
Biosciences) in 20 mM HEPES, pH 7.0, and incubated with a
stoichiometric excess of nickel chloride. For the removal of the excess
nickel(II) ions, an additional step of anion exchange chromatography
(15) Rao, J.; Lahiri, J.; Isaacs, L.; Weis, R. M.; Whitesides, G. M. Science 1998,
280, 708-711.
(16) Kitov, P. I.; Sadowska, J. M.; Mulvey, G.; Armstrong, G. D.; Ling, H.;
Pannu, N. S.; Read, R. J.; Bundle, D. R. Nature 2000, 403, 669-672.
(17) Lee, R. T.; Lee, Y. C. Glycoconjugate J. 2000, 17, 543-551.
(18) Mourez, M.; Kane, R. S.; Mogridge, J.; Metallo, S.; Deschatelets, P.;
Sellman, B. R.; Whitesides, G. M.; Collier, R. J. Nat. Biotechnol. 2001,
19, 958-961.
(19) Lundquist, J. J.; Toone, E. J. Chem. ReV. 2002, 102, 555-578.
(20) Zhang, Z.; Merritt, E. A.; Ahn, M.; Roach, C.; Hou, Z.; Verlinde, C. L.;
Hol, W. G.; Fan, E. J. Am. Chem. Soc. 2002, 124, 12991-12998.
(21) Christensen, T.; Gooden, D. M.; Kung, J. E.; Toone, E. J. J. Am. Chem.
Soc. 2003, 125, 7357-7366.
(22) Mulder, A.; Auletta, T.; Sartori, A.; Del Ciotto, S.; Casnati, A.; Ungaro,
R.; Huskens, J.; Reinhoudt, D. N. J. Am. Chem. Soc. 2004, 126, 6627-
6636.
(23) Benito, J. M.; Gomez-Garcia, M.; Ortiz Mellet, C.; Baussanne, I.; Defaye,
J.; Garcia Fernandez, J. M. J. Am. Chem. Soc. 2004, 126, 10355-10363.
(24) Hochuli, E.; Dobeli, H.; Schacher, A. J. Chromatogr. 1987, 411, 177-
184.
(25) Ueda, E. K.; Gout, P. W.; Morganti, L. J. Chromatogr., A 2003, 988, 1-23.
(26) Hart, C.; Schulenberg, B.; Diwu, Z.; Leung, W. Y.; Patton, W. F.
Electrophoresis 2003, 24, 599-610.
(35) Dorn, I. T.; Neumaier, K. R.; Tampe, R. J. Am. Chem. Soc. 1998, 120,
2753-2763.
(27) Sigal, G. B.; Bamdad, C.; Barberis, A.; Strominger, J.; Whitesides, G. M.
Anal. Chem. 1996, 68, 490-497.
(36) Kapanidis, A. N.; Ebright, Y. W.; Ebright, R. H. J. Am. Chem. Soc. 2001,
123, 12123-12125.
(28) Gershon, P. D.; Khilko, S. J. Immunol. Methods 1995, 183, 65-76.
(29) Schmid, E. L.; Keller, T. A.; Dienes, Z.; Vogel, H. Anal. Chem. 1997, 69,
1979-1985.
(37) Hutschenreiter, S.; Neumann, L.; Radler, U.; Schmitt, L.; Tampe, R.
ChemBioChem 2003, 4, 1340-1344.
(30) Xu, C.; Xu, K.; Gu, H.; Zhong, X.; Guo, Z.; Zheng, R.; Zhang, X.; Xu, B.
J. Am. Chem. Soc. 2004, 126, 3392-3393.
(38) Guignet, E. G.; Hovius, R.; Vogel, H. Nat. Biotechnol. 2004, 22, 440-
444.
(31) Schmitt, L.; Ludwig, M.; Gaub, H. E.; Tampe, R. Biophys. J. 2000, 78,
3275-3285.
(39) Nieba, L.; Nieba-Axmann, S. E.; Persson, A.; Hamalainen, M.; Edebratt,
F.; Hansson, A.; Lidholm, J.; Magnusson, K.; Karlsson, A. F.; Pluckthun,
A. Anal. Biochem. 1997, 252, 217-228.
(32) Schmitt, L.; Dietrich, C.; Tampe, R. J. Am. Chem. Soc. 1994, 116, 8485-
8491.
(40) Dorn, I. T.; Pawlitschko, K.; Pettinger, S. C.; Tampe, R. Biol. Chem. 1998,
379, 1151-1159.
(33) Dietrich, C.; Schmitt, L.; Tampe, R. Proc. Natl. Acad. Sci. U.S.A. 1995,
92, 9014-9018.
(41) Johnson, R. D.; Todd, R. J.; Arnold, F. H. J. Chromatogr., A 1996, 725,
225-235.
(34) Dorn, I. T.; Eschrich, R.; Seemuller, E.; Guckenberger, R.; Tampe, R. J.
Mol. Biol. 1999, 288, 1027-1036.
(42) Lata, S.; Piehler, J. Anal. Chem. 2005, 77, 1096-1105.
9
10206 J. AM. CHEM. SOC. VOL. 127, NO. 29, 2005

Adaptors for Recognition of Histidine-Tagged Proteins
A R T I C L E S
Scheme 1. Synthesis of the Chelator Building Blocks and the Multivalent Chelator Headgroupsa
a (A) mono-NTA, as well as bis-NTA and tetrakis-NTA obtained by a dendrimer approach. (B) tris-NTA derived from a cyclic scaffold.
under the same conditions was carried out. The structures of the
fluorescein-conjugated chelator heads after nickel(II) ion loading (20,
21, 22, and 23), used in all the assays to follow, are shown in Figure
1.
oligohistidine-tagged ifnar2-EC (5 µM) in 20 mM HEPES, pH 7.5,
and 150 mM NaCl (HEPES-buffered saline, HBS) for 20 min. The
sample was then loaded onto an SEC column (Superdex 200 PC 3.2/
30) in a SMART system (Amersham Biosciences), and elution was
monitored at 280 and 490 nm. To determine the relative amount of
protein, the absorbance at 280 nm was corrected by subtracting the
contribution of the fluorescence dye, which was determined from the
absorbance at 490 nm. The chromatograms were normalized to the
corrected absorption at 280 nm. Thus, the relative height of the 490-
nm signal is a measure of labeling efficiency and hence complex
stability. A 280 nm:490 nm ratio of ∼1:2.5 was expected for a 1:1
complex.
Isothermal Titration Calorimetry. Calorimetric measurements were
carried out using a VP-ITC instrument with a cell volume of 1.4115
mL (MicroCal, LLC). For ITC measurements, the proteins were purified
in the same batch of the stock buffer as used for dilution of the sample
and injectant. Exclusively one single batch of buffer was used for
sample preparation and dilution for the chelator heads as well as the
proteins. The background heat changes were negligible under these
conditions. Interaction constants characterizing the chelator head
Protein Expression and Purification. Wild-type extracellular
domain of the human type I interferon receptor subunit ifnar2 (ifnar2-
EC) variants with a C-terminal hexahistidine tag (ifnar2-H6) and a
C-terminal decahistidine-tag (ifnar2-H10) were expressed and purified
as reported earlier.43 Maltose binding protein (MBP) with a C-terminal
decahistidine and a hexahistidine tag (MBP-H10 and MBP-H6,
respectively) were expressed and purified as previously described.42
To check the integrity of the histidine tag, the purified proteins were
analytically digested with factor Xa, and the molecular mass of the
cleaved peptide was confirmed by MALDI-MS. All purified proteins
were more than 95% homogeneous and monomeric as detected by
nonreducing SDS-PAGE and size exclusion chromatography (SEC).
Analytical Size Exclusion Chromatography. Two-fold molar
excess of the chelator heads (10 µM) was incubated with the
(43) Piehler, J.; Schreiber, G. J. Mol. Biol. 1999, 289, 57-67.
9
J. AM. CHEM. SOC. VOL. 127, NO. 29, 2005 10207

A R T I C L E S
Lata et al.
Figure 1. Chemical structures of the carboxyfluorescein-conjugated chelator heads: mono-NTA-Fluo (20), bis-NTA-Fluo (21), tris-NTA-Fluo (22), and
tetrakis-NTA-Fluo (23). The conjugates are shown loaded with nickel(II) ions as used for interaction studies in solution. The two remaining coordination
sites of the NTA-complexed nickel(II) ions are occupied by general ligands X, and possible charges are therefore omitted.
oligohistidine tag interactions were determined by direct titration at
pH 7.5 in HBS at 25 °C. The interaction assessments were carried out
by adding the tagged protein or peptides to the chelator head solution
in the cell. The solution in the cell was stirred at 800 rpm by the syringe
to ensure rapid mixing. Typically 4-10 µL of the injectant was added
over 30 s with an adequate interval between the injections to allow
complete equilibration, and the heat change was monitored continuously
upon addition of each injection of the titrant as a function of time. The
data were analyzed using the Origin software package supplied by
MicroCal. An interaction model assuming n independent and equivalent
binding sites was applied, and the stoichiometry n, change in enthalpy
∆H°, and binding constant KD were iteratively fitted.
Time-Resolved Fluorescence Measurements. The kinetics of the
chelator/oligohistidine tag complex formation and dissociation were
measured using carboxyfluorescein-labeled H6 and H10 peptides. Upon
complex formation of these peptides with the Ni2+-loaded chelator/
carboxyfluorescein conjugates, strong fluorescence quenching was
the arrows.
observed (Figure 2), which was probably mainly due to the chelated
transition metal ions.37,44,45 This effect was used as a readout for
Figure 2. Schematic representation of bis-NTA coupled to fluorescein (A)
and its complex with fluorescein-labeled H6 (B). The quenching processes
leading to fluorescence decay upon complex formation are indicated by
following the interaction in real time.
H10. MBP-H10 (5 µM) was added to an equilibrated complex (500
nM each) of a given interaction pair in a cuvette, and the fluorescence
at 520 nm was acquired as a function of time for up to 1 h. The
equilibrium signal was determined after overnight incubation. For
evaluation and comparison of the curves, the fluorescence signal at t0
was subtracted and the curves were inverted and normalized to the
equilibrium signal amplitude. The dissociation rate constant kd was
determined from these curves by fitting a first-order kinetics model:
Association kinetics for all the chelator heads and the dissociation
kinetics for mono-NTA were measured at 25 °C using an Applied
Photophysics (π* 180 CDF) stopped flow spectrometer equipped with
a photomultiplier tube and thermostat. The fluorophores were excited
at 490 nm, and the signal was collected through a 515-nm long pass
filter using a photomultiplier tube. The dissociation kinetics for the
MCHs were measured with a fluorescence spectrometer (Varian Cary
Eclipse) by exciting the fluorophores at 470 nm. Depending upon the
kinetics of the process, single wavelength (510 nm) or spectral (490-
700 nm) data acquisition was carried out. Kinetics were fitted using
Origin (Microcal Software) or Biaevaluation 2.1 (Biacore) for explicit
models, and Berkeley Madonna (UCB) for numerical integration. For
MCH/oligohistidine tag interactions, the dissociation kinetics were
measured by chasing the fluorescein-labeled tag with nonlabeled MBP-
S(t) ) S0‚exp[-kd‚(t - t0)]
(1)
where S(t) is the signal at time t, and S0 is the equilibrium signal
amplitude. The association kinetics for MCHs with oligohistidine were
measured by mixing equal volumes of the chelator head (10 µM) and
the oligohistidine tag (10 µM), leading to a final concentration of 5
µM of each. Since the association under these conditions was much
faster than dissociation for all MCH/oligohistidine pairs, the curves
(44) Kemlo, J. A.; Shepherd, T. M. Chem. Phys. Lett. 1977, 47, 158-162.
(45) Fabbrizzi, L.; Licchelli, M.; Pallavicini, P.; Sacchi, D.; Taglietti, A. Analyst
1996, 121, 1763-1768.
9
10208 J. AM. CHEM. SOC. VOL. 127, NO. 29, 2005

Adaptors for Recognition of Histidine-Tagged Proteins
A R T I C L E S
Table 1. Binding Parameters of the Chelator/Oligohistidine Complexes
mono-NTA
bis-NTA
tris-NTA
tetrakis-NTA
H6
H10
H6
H10
H6
H10
H6
H10
m/ra
2/4
2/8
4/2
4/6
6/0
∼1
1
20 ( 10
-35.0 ( 1.2
6.3
6/4
6/2
∼1
1
40 ( 20
-37.5 ( 1.3
5.7
61
20 ( 4
2.7 ( 0.8
37 ( 10
-9.9 ( 0.2
-93 ( 3
8/2
∼1
1
80 ( 30
-50.6 ( 1.5
4.0
Γ ( )b
∼0
∼0
∼0.15
∼1
∼1
n ( )c
0.35
-
1
1
1
K
D (nM)d
(14 ( 4) × 103
-40.9 ( 1.5
121
-
270 ( 50
-25.0 ( 1.0
5.8
240 ( 50
-27.9 ( 1.1
4.3
40 ( 20
-41.4 ( 1.5
4.5
∆H° (kcal/mol)e
c (cell) (µM)f
-
-
c (syr.) (µM)f
789
-
61
100
61
100
100
kd (10-3 s-1
)
1800 ( 400
1.4 ( 0.4
(13 ( 5) × 103
-6.5 ( 0.2
-24 ( 2
1400 ( 300
2.7 ( 0.7
25 ( 4
1.5 ( 0.4
68 ( 20
-9.5 ( 0.2
-52 ( 2
3 ( 0.5
3.1 ( 0.8
10 ( 3
-10.7 ( 0.3
-58 ( 2
0.34 ( 0.01
1.6 ( 0.4
2.1 ( 0.8
-11.6 ( 0.3
-79 ( 3
∼0.05
0.55 ( 0.1
4.6 ( 1.4
1.2 ( 0.4
-11.9 ( 0.3
-130 ( 4
g
ka (105 M-1 s-1
D (nM)i
∆G° (kcal/mol)j
∆S° (cal/mol/K)k
)
3.5 ( 1
h
K
(5 ( 2) × 103
-7.1 ( 0.2
-22 ( 1
0.14 ( 0.05
-13.1 ( 0.3
-95 ( 3
a Theoretical multivalency (m) and redundancy (r) of the interaction pairs (1:1 stoichiometry). b Labeling degree Γ estimated from analytical SEC. c Number
of chelator heads per oligohistidine-tagged protein. d Equilibrium dissociation constant KD determined by ITC. e Binding enthalpy ∆H° for a 1:1 interaction.
f Concentrations in the cell (chelator head) and in the syringe (oligohistidine-tagged protein). g Dissociation rate constant kd determined by time-resolved
fluorescence measurements. h Association rate constant ka determined by time-resolved fluorescence measurements. i Equilibrium dissociation constant KD
determined from ka and kd. j ∆G° determined from KD obtained from kinetic data. k ∆S° determined from ∆G° (kinetic data) and ∆H° (ITC data).
were fitted by a model for irreversible second-order association kinetics:
relatively constrained due to the shorter chain length of synthon
12 compared to that of 3 and the cyclic scaffold. For probing
the interaction with oligohistidine tags, carboxyfluorescein was
coupled to the primary amino group of the chelator heads as a
versatile and sensitive fluorescence reporter. The removal of
excess nickel(II) ions after quantitatively loading the NTA
groups without affecting the chelated metal ions was a critical
issue, which was resolved by anion exchange chromatography
at neutral pH. The attached carboxyfluorescein was beneficial
for purification of the chelator heads after loading with metal
ions and for determining the chelator concentrations photo-
metrically. All measurements were carried out with the Ni(II)
ion loaded fluorescein conjugates (Figure 1), which will
nonetheless be referred to by their chelator heads (mono-NTA,
bis-NTA, tris-NTA, and tetrakis-NTA).
ka(t - t0)‚c2
S(t) ) S0‚
(2)
1 + ka(t - t0)‚c
where ka is the association rate constant, and c is the concentration of
the reactants. The association kinetics were measured for the mono-
NTA/oligohistidine tag interaction as described for the MCH. For
measuring dissociation kinetics, a pre-equilibrated, stoichiometric
mixture of mono-NTA (5 µM) and the oligohistidine tags (5 µM) was
diluted 11 times. These curves were evaluated by fitting the differential
equation for a reversible second-order reaction:
dS(t)
dt
) ka[S0 - S(t)]2 - kd[S(t)]
(3)
Details of the fitting procedure are described in Supporting Information.
The equilibrium dissociation constant KD was determined from the
rate constants:
To exploit the maximum conformational space of a hexa-
and a decapeptide, we chose cumulated H6 and H10 tags for
testing the chelator heads. The chelator heads presented 1-4
metal ions per molecule, leading to 2-8 possible coordinative
bonds. On the other hand, H6 and H10 had the overall potential
of 6 and 10 coordinations, respectively. The number of possible
coordinative bonds (termed multivalency in the following) varied
between the different chelator head/oligohistidine pairs. Thus,
theoretically the number of histidines on the tag either over- or
under-fulfill the coordination demands of the chelators in most
instances with a single exception of tris-NTA hexahistidine
interaction. This adds a new feature to the interaction, which
we refer to as redundancy in the following. The theoretical
multivalencies and redundancies in terms of possible coordina-
tion bonds are listed in Table 1 for every given interaction pair.
The targets carrying the oligohistidine tags were two different
monomeric proteins and one fluorophore, all linked to the
N-terminus of the tag: (i) the ifnar2-EC, (ii) MBP, and (iii)
carboxyfluorescein attached through an aminohexanoic acid
linker.
KD ) kd/ka
(4)
The standard free energies ∆G° with respect to 1 M concentration were
calculated from the equilibrium dissociation constant KD according to
eq 5:
∆G° ) RT ln(KD/1 M)
(5)
and the interaction entropies were determined from these ∆G° and the
∆H° values from ITC measurements based on eq 6:
∆G° ) ∆H° - T∆S°
Results and Discussion
(6)
Grafting Metal-Chelating Moieties onto Branched and
Cyclic Scaffolds. The MCHs were built up using tert-butyl-
protected NTA synthons (3, 12), which were grafted onto
different scaffolds. The modular synthesis enabled versatile use
of different scaffolds and is in principle also compatible with
solid-phase synthesis. MCHs with two, three, and four NTA
moieties were obtained, which were called bis-NTA (17), tris-
NTA (19), and tetrakis-NTA (18), respectively. These MCHs
are based on different types of scaffolds: while bis-NTA and
tetrakis-NTA are based on a dendritic architecture with rather
high conformational flexibility, tris-NTA is conformationally
Stable and Stoichimetrically Defined MCH/Oligohistidine
Complexes. Binding of the Ni2+-loaded chelator/fluorescein
conjugates to H6- and H10-tagged ifnar2-EC (ifnar2-H6 and
ifnar2-H10, respectively) was investigated by analytical SEC.
The protein was equilibrated with a 1.5-fold stoichiometric
excess of the chelator head, and mixture was separated by SEC.
9
J. AM. CHEM. SOC. VOL. 127, NO. 29, 2005 10209

A R T I C L E S
Lata et al.
Figure 3. Stability and stoichiometry of chelator/oligohistidine complexes studied by analytical SEC. The absorbance was monitored at 280 nm (black) and
490 nm (blue) to discriminate the chelator/fluorescein conjugate and the protein. (A-E) ∼1.5-fold molar excess of the chelator/fluorescein conjugates. (A)
Full chromatogram for tris-NTA/ifnar2-H6 showing the exclusion volume (VE), the bed volume (VB), the protein peak (I), and the peak of the chelator/
fluorophor conjugate (II). (B-E) Comparison of the complexes with ifnar2-H6 (left panel) and ifnar2-H10 (right panel). (B) mono-NTA. (C) bis-NTA. (D)
tris-NTA. (E) tetrakis-NTA. In C-E, the elution profile of the free protein is shown for comparison (orange). In the left panel of C, the protein peak of the
right panel is shown (- - -). (F,G) SEC with ∼2-fold molar excess of the protein. (F) Full chromatogram for tetrakis-NTA with a stoichiometric excess
of ifnar2-H6. (G) Protein peaks for bis-NTA/ifnar2-H10 (left panel) and tetrakis-NTA/ifnar2-H6 (right panel) complexes in comparison to the peak for
ifnar2-EC alone (orange). For visual comparison, the peaks are shown in arbitrary units (a.u.).
A typical chromatogram with simultaneous detection at 280 and
490 nm is shown in Figure 3A, clearly showing one peak for
the protein/chelator complex and one for the free dye eluting
with the bed volume.
comparison of ifnar2-H6 was seen. Unlike ifnar2-H6, the
chelator head and the protein peaks overlapped perfectly,
suggesting that the bis-NTA/ifnar2-H10 complex remained
stable during SEC. The increase in labeling efficiency was
accompanied by a nearly quantitative shift of the protein peak
toward higher apparent molecular weight, confirming a labeling
efficiency of >90% for this interaction pair. Given four
coordination sites on bis-NTA, both H6 and H10 over-satisfy
the coordination demands, although by different redundancy
factors. Thus, higher redundancy of the H10 tag significantly
increased binding affinity toward this MCH.
The relative absorbance signals at these two wavelengths
within the protein peak, as well as the peak position, were used
to estimate the amount of fluorescence dye attached to the
protein. Shown in Figure 3B-E are the elution peaks of ifnar2-
H6 and ifnar2-H10 with the different chelator heads measured
under the same conditions. When ifnar2-H6 was incubated with
mono-NTA, retention of the chelator head with the protein peak
was hardly detectable (Figure 3B). A similar observation was
made with ifnar2-H10 as well, but a slight increase of the
labeling degree was observed. In contrast, when ifnar2-H6 was
incubated with bis-NTA, the labeling efficiency substantially
increased (Figure 3C), but the chelator head-to-protein ratio was
still only 10-15%. At the same time, a minor shift of the
chelator head peak relative to that of the protein peak was
observed, suggesting dissociation of the complex during the SEC
experiment. Interestingly, when ifnar2-H10 was incubated with
bis-NTA, nearly 10-fold increase in labeling efficiency in
When ifnar2-H6 was incubated with tris- and tetrakis-NTA,
high labeling efficiencies (>90%) were observed, and the protein
and the chelator head peaks perfectly coincided (Figure 3D,E).
A shift of the protein peak was observed for tris-NTA, which
was similar to the shift observed for the bis-NTA/ifnar2-H10
complex. For tetrakis-NTA, an even stronger shift of the
complexed protein was observed (Figure 3E). No free protein
was detectable, confirming labeling efficiencies close to 100%
for these interaction pairs. These observations impressively
confirmed the expected effect of increasing complex stability
9
10210 J. AM. CHEM. SOC. VOL. 127, NO. 29, 2005

Adaptors for Recognition of Histidine-Tagged Proteins
A R T I C L E S
Figure 4. ITC thermograms (top) and titration curves (bottom) of the chelators titrated with MBP-H6. (A) mono-NTA. (B) bis-NTA. (C) tris-NTA. (D)
tetrakis-NTA. The line corresponds to a fit of a standard interaction model assuming a single type of binding sites.
with increasing number of NTA groups (multivalency). Very
similar chromatograms were obtained for the interaction of
tetrakis-NTA and tris-NTA with ifnar-H10. The dye-to-protein
ratio of the SEC chromatogram at stoichiometric excess of MCH
already suggested a 1:1 stoichiometry of the complex. Stable
and stoichiometric complex formation was further corroborated
by SEC experiments with substoichiometric amounts of the
MCH (Figure 3F,G). No free MCH was detectable under these
conditions (Figure 3F). The protein peak was broadened as
detected at 280 nm; strikingly, however, the peak observed at
490 nm was still eluting with the same shape and at the same
elution volume as observed for stoichiometric excess of MCH
(Figure 3G). Thus, a stable 1:1 interaction was confirmed for
bis-NTA/H10, as well as for tris-NTA and tetrakis-NTA with
both H6 and H10.
Similar labeling efficiencies were obtained with MBP-H6 and
MBP-H10 as with ifnar2-H6 and ifnar2-H10 (data not shown),
substantiating generic recognition of oligohistidine-tagged pro-
teins by the MCH. The interaction was entirely specific for
oligohistidine tags: when tagless ifnar2-EC or MBP was
incubated with tris-NTA, no dye was detectable within the
protein peak (data not shown). The interaction was fully
reversible, as no complex was detectable after addition of 50
mM EDTA (data not shown).
titration for mono-NTA (121 µM) with MBP-H6 (790 µM).
While at these high concentrations the thermograms showed
very large changes in heat, the curvature of the binding isotherm
implies low affinity. Table 1 summarizes the enthalpy ∆H°,
interaction stoichiometry n, and equilibrium dissociation constant
KD, as obtained by fitting to the binding isotherm a 1:n
interaction model, which assumes a single type of binding sites.
Strikingly, a stoichiometric ratio of three mono-NTA units per
MBP-H6 was observed, suggesting that all six histidines were
coordinated under these conditions. A total ∆H° of 40.9 kcal/
mol was obtained, implying that the average heat of coordination
per mono-NTA was 13.6 kcal/mol.
This titration yielded a KD of 14 ( 4 µM, which is in good
agreement to what has been previously measured for mono-
NTA/MBP-H6 interaction in solution.35,36,38 The MCHs were
titrated at much lower concentrations, and the clear sigmoidal
shape of the binding isotherms corroborated affinities substan-
tially higher compared to that of mono-NTA. Figure 4B shows
the binding isotherm when bis-NTA (5.8 µM) was titrated with
MBP-H6 (60.7 µM). The interaction parameters are listed in
Table 1. A clear 1:1 stoichiometry was observed at the employed
concentration, with a KD of 270 nM. Thus, a 50-fold increase
in affinity compared to mono-NTA was achieved. The ∆H° of
25 kcal/mol is nearly twice as much as that obtained for mono-
NTA. When tris-NTA and tetrakis-NTA were titrated with
MBP-H6, under conditions similar to bis-NTA, an even nar-
rower transition of the binding isotherm was observed, indicating
a further increase in affinity (Figure 4C,D). The stoichiometries
were 1:1, and the KD values determined for tris-NTA and
tetrakis-NTA were 23 and 41 nM, respectively. Compared to
that of bis-NTA, a substantial increment in ∆H° was observed
for tris-NTA. However, compared to that of tris-NTA, no further
increase in ∆H° was seen for tetrakis-NTA. This is in line with
the fact that MBP-H6 allows formation of a maximum of six
coordinative bonds for both tris-NTA and tetrakis-NTA.
Additive ∆H° at High Redundancies. All three MCHs were
also titrated with MBP-H10, and similar titration curves (not
shown) were obtained as for MBP-H6. These data are sum-
marized in Table 1. In all the cases, 1:1 interactions were
concluded from the fits. For a given MCH, a systematic increase
Large Gains in ∆H° with Increasing Multivalency. ITC
was performed to quantitatively assess and segregate the
contributions of enthalpy and entropy toward the affinity
constants for the chelator head binding to the oligohistidine tags.
To understand the additive effects of increasing number of
coordinations arising from the multivalent nature of the binding
partners, the chelator heads were titrated with H6- and H10-
tagged proteins and peptides. Typical titrations for all the
chelator heads with MBP-H6 are compared in Figure 4. The
complex formation was always net exothermic as would be
expected for a reaction in which the coordination bonds are
formed, but very different curvatures of the titration curves were
observed depending upon the chelator head. To achieve suf-
ficient curvature in the binding isotherm, titrations with mono-
NTA were carried out at concentrations substantially higher than
the concentrations used with the MCH. Figure 4A shows the
9
J. AM. CHEM. SOC. VOL. 127, NO. 29, 2005 10211

A R T I C L E S
Lata et al.
Figure 5. ∆H° of the chelator/oligohistidine pairs as obtained by ITC. (A) Correlation of the molar binding enthalpies ∆H° obtained for MBP-H6 (O) and
MBP-H10 (9) with the number of potential coordination bonds (multivalency). (B) Difference ∆∆Hc-m° between the measured and the theoretical, additive
∆H° for different multivalencies separated into different redundancies. The average ∆∆Hc-m° values for different redundancies are shown as gray bars in
the background.
in ∆H° was observed for MBP-H10 compared to that of MBP-
H6. Moreover, this difference (∆∆HH10-H6°) increased with
increasing multivalency. The KD values for MBP-H10 with a
given MCH, however, did not substantially differ as compared
to those for MBP-H6 and even appeared to be systematically
higher than those for the corresponding MCH/H6 complexes.
This was surprising, as SEC conclusively indicated substantially
increased affinity for MBP-H10 as compared to that for MBP-
H6, at least for bis-NTA. Therefore, the binding affinities were
independently determined by measuring the interaction kinetics
(see below).
oligohistidine complexes in a quantitative manner, association
and dissociation kinetics were determined. All rate constants
and equilibrium constants obtained from these experiments are
summarized in Table 1. The change of fluorescence with time
when mono-NTA (5 µM) was mixed with a carboxyfluorescein-
labeled H6 peptide (5 µM) in a stopped flow setup is shown in
Figure 6A. For analyzing the dissociation kinetics, the equili-
brated complex (5 µM of each) was diluted 11-fold, and the
emission was followed in real time (Figure 6B). The second-
order kinetics for a 1:1 interaction (eq 3) with the appropriate
initial conditions was iteratively fitted to the mixing and the
dilution phases until the association and the dissociation rate
constants converged. For the mono-NTA interaction with H6,
a kd of 1.8 s-1 and a ka of 1.4 × 105 M-1 s-1 were determined.
Using eq 4, we calculated a KD value of 13 µM from these rate
constants, which is in good agreement to the Kd obtained by
ITC. Likewise, the kinetic constants were determined for H10.
The mixing and the dilution phases are shown in Figure 6C,D.
In comparison to H6, an approximately 2-fold higher ka and a
slight reduction in kd were observed. The determined kd and ka
were 1.4 s-1 and 2.7 × 105 M-1 s-1, respectively, yielding a
KD of 5 µM. Thus, only a minor increase in affinity from H6 to
H10 was observed, which contradicts the considerable differ-
ences observed in complex stability between H6 and H10 tags
with immobilized chelators.42
Figure 5A shows the ∆H° for all MCH/oligohistidine
complexes as a function of theoretical number of coordination
bonds.
Assuming a ∆H° of 13.6 kcal/mol (the ∆H° for the mono-
NTA/MBP-H6 interaction) as the ∆H° for an NTA moiety
coordinated by two histidines, the straight line represents the
additive increase in heat as a function of number of coordinative
bonds, which coincides well overall with the experimental data.
The highest deviation from the theoretically expected ∆H° was
observed for tris-NTA/H6, in agreement with the lowest
redundancy of this interaction pair. The theoretical values for
∆H° were reached for bis-NTA/H10 and nearly reached for tris-
NTA/H10. The difference ∆∆Hc-m° between measured ∆H°
and calculated ∆H° is compared in Figure 5B for different
multivalencies as a function of redundancy. This comparison
reveals two striking features of multivalency and redundancy
on ∆H°: (i) the theoretical ∆H° values are generally reached
closer with increasing redundancy. For redundancies of 4 and
higher, the ∆∆Hc-m° were negligible. (ii) For pairs with the
same redundancy, the ∆∆Hc-m° increased with increasing
multivalency. Both effects can be ascribed to steric constraints
upon complex formation:46 high redundancy increases the
possibility to adopt a complex, in which all coordination sites
are occupied by histidines at a minimum strain; high multi-
valency increases the number of constraints, which have to be
fulfilled to occupy all coordination sites simultaneously.
Increase in Complex Stability by Several Orders of
Magnitude. To assess affinities and stabilities of the chelator/
For the MCH/oligohistidine complexes, no significant dis-
sociation was observed upon dilution, which is in line with the
substantially higher stability of these complexes observed in
SEC. Therefore, the dissociation rate constants were determined
by adding a 10-fold molar excess (5 µM) of MBP-H10 to an
equilibrated complex of the chelator head and the labeled peptide
(500 nM of each). Typical dissociation phases after correction
and normalization are shown in Figure 7A. First-order kinetics
was assumed under these conditions, and the dissociation rate
constants were determined by fitting eq 1. The dissociation rate
constants for the different chelator head/oligohistidine pairs are
compared in Figure 7B. A decrease in the kd of more than 4
orders in magnitude was observed from mono-NTA/H10 to tris-
NTA/H10. With increasing number of NTA moieties of the
MCH, the relative difference in complex stability between H6
and H10 increased. The most surprising feature, however, was
(46) Mammen, M.; Choi, S. K.; Whitesides, G. M. Angew. Chem., Int. Ed. 1998,
37, 2755-2794.
9
10212 J. AM. CHEM. SOC. VOL. 127, NO. 29, 2005

Adaptors for Recognition of Histidine-Tagged Proteins
A R T I C L E S
Figure 6. Kinetics of the interaction of mono-NTA with H6 (A,B) and H10 (C,D) as monitored by stopped-flow fluorescence. (A,C) Association kinetics.
(B,D) Dissociation kinetics. The fit of the model is shown as a red curve.
Figure 7. Dissociation kinetics of the MCH/oligohistidine complexes as characterized by fluorescence dequenching upon complex dissociation. (A) Dissociation
kinetics for different MCHs in complex with H6 (‚‚‚) and H10 (s) peptides as observed by fluorescence dequenching upon adding 10 µM unlabeled
MBP-H10 (black: bis-NTA; blue: tris-NTA; red: tetrakis-NTA). The curves were processed as described in the Experimental Section. (B) Comparison of
the kd values for complexes of MCH with H6 (gray) and H10 (white) peptides obtained by a monoexponential fit of the dissociation curves as shown in B
compared to the kd obtained for mono-NTA. (C) Equilibrium dissociation constants as determined from the interaction kinetics in comparison for H6 (gray)
and H10 (white).
the very low stability of the tetrakis-NTA/H6 complex, which
dissociated nearly as fast as the bis-NTA/H6 complex. This is
in marked contrast to SEC experiments, where the tetrakis-NTA/
ifnar2-H6 complex showed substantially higher stability as
compared to the bis-NTA/H6 complex. The association rate
constants for the MCHs were determined by stopped-flow
fluorescence as for mono-NTA. The association rate constants
therefore were determined by fitting eq 2, as the contribution
of complex dissociation to the curvature was negligible. The
values are listed in Table 1.
The differences in the ka for the different pairs were small
overall. However, association of H10 was always faster than
association of H6. This could be explained by a higher
probability of a successful collision: the greater the number of
histidines the higher the probability of encounter complex
formation. This effect is not as pronounced for the increasing
number of NTA moieties from mono- to tetrakis-NTA. Two
opposing factors might be at work: (i) association rate enhance-
ment in proportion to the number of chelation sites and (ii)
electrostatic repulsion resulting from the greater number of
chelator heads carrying negative charges leading to more
repulsion and hence a reduction in association rate.
High Loss in Entropy for Multivalent Complexes. We
employed three different approaches for determining the stability
and affinity of the chelator head oligohistidine tag interaction.
The equilibrium dissociation constants (KD) for all the chelator/
oligohistidine pairs as calculated from the rate constants
according to eq 4 are compared in Figure 7C. Table 1
summarizes the KD obtained from ITC and the labeling
efficiencies as estimated from SEC. Both ITC and kinetic
measurements yielded very similar KD values for mono-NTA/
H6 and bis-NTA/H6 interaction. In contrast, the KD values
determined by ITC as compared to kinetic analysis were much
higher for the bis-NTA/H10, tris-NTA/H6, tris-NTA/H10, and
9
J. AM. CHEM. SOC. VOL. 127, NO. 29, 2005 10213

A R T I C L E S
Lata et al.
Figure 8. Schematic of different scenarios of multivalency and redundancy. (A) Redundancy of the oligohistidine tag (e.g., bis-NTA/H6) guarantees maximum
∆G°, but allows nonstoichiometric complexes. (B) High multivalency and no redundancy as in the case of tris-NTA/H6 yields stable, stoichiometric complexes.
(C) Redundancy of chelator groups as in the tetrakis-NTA/H6 complex provides docking sites for active exchange.
Figure 9. (A) Comparison of ∆H° and ∆G° and T∆S° for all chelator complexes with H6 (gray) and H10 (white). (B) Correlation of ∆G° with ∆H° (i.e.,
effective multivalency) for the different chelator complexes with H6 (9) and with H10 (O). The points obtained for each MCH are linked by a line. The
dotted line indicates additive ∆G° of a noncooperative multivalent interaction as defined by Whitesides and co-workers.46
tetrakis-NTA/H10 complexes. These discrepancies could arise
from the experimental specifics related to the multivalent and
redundant nature of the interaction partners:47
SEC and ITC measurements. We interpret this apparently low
complex stability by invoking nonspontaneous dissociation: H6
bound to tetrakis-NTA was actively exchanged by the com-
peting MBP-H10 due to the additional, noncomplexed NTA
moiety, which can act as a docking site for a competitor (Figure
8C).
(i) The KD values determined by ITC are orders of magnitude
below the concentration applied in these measurements. Under
these conditions, a titration curve contains only minute informa-
tion about the affinity. At the same time, transient formation of
the complexes with stoichiometries other than 1:1 due to a high
concentration (micromolar) and high excess of the MCH in the
cells cannot be completely ruled out (cf. Figure 8A), which may
in turn slightly bias the binding isotherm. Therefore, extracting
reliable affinity information from these measurements for the
MCH is not feasible. However, in some instances the relative
affinities can be conclusively determined (e.g., a significant
increase in steepness of the binding isotherm for tetrakis-NTA/
H6 in comparison with that of bis-NTA/H6 clearly indicates a
substantial decrease in KD from bis-NTA/H6 to tetrakis-NTA/
H6 complex). This observation is also in good agreement with
the labeling efficiencies as obtained from SEC experiment.
(ii) The stability of the tetrakis-NTA/H6 complex as deter-
mined by the competition assay contradicts the relatively higher
stability and affinity compared to bis-NTA/H6 as is evident from
This would be an interesting mechanism based on redun-
dancy, which could be exploited for applications that require
fast exchange of the tag bound to the MCH by another tag.
Since this effect is probably a specific feature of the tetrakis-
NTA/H6 complex, the KD values obtained from kinetic mea-
surements appeared generally more reliable compared to the
KD obtained from ITC. Therefore, the free energies ∆G° of
complex formation were determined from the KD obtained from
kinetic data using eq 5. These ∆G° values are shown in
comparison to the ∆H° and the resulting T∆S° for all the pairs
in Figure 9A. In all the cases, ∆H° was substantially higher
than ∆G° (i.e., the ∆S° was negative). For a given histidine
tag, the ∆S° increased with increasing number of NTA moieties,
indicating an increasing “penalty” with the number of coordina-
tion bonds being formed. This is not surprising for a multivalent
interaction between two flexible binding partners, both of which
substantially lose conformational freedom upon complex forma-
tion.
(47) Kitov, P. I.; Bundle, D. R. J. Am. Chem. Soc. 2003, 125, 16271-16284.
9
10214 J. AM. CHEM. SOC. VOL. 127, NO. 29, 2005

Adaptors for Recognition of Histidine-Tagged Proteins
A R T I C L E S
The correlations of ∆G° with multivalency and with ∆H°
are shown in Figure 9B. The additive increase of ∆G° with
multivalency of a noncooperative multivalent interaction (as
defined by Whitesides and co-workers)46,48 is indicated by a
dotted line. For all MCH/oligohistidine pairs, substantially less
than additive ∆G° was obtained, that is, negative cooperativity,
which is typical for multivalent interactions.46 The good
correlation of ∆G° with ∆H° supports that complex formation
is mostly driven by enthalpy even at high redundancies (Figure
9B): a constant increase was observed for mono-, bis-, and tris-
NTA, which did not significantly differ for H6 and H10. This
correlation did not hold true for the tetrakis-NTA/oligohistidine
complexes, for which significantly lower ∆G°/∆H° ratios were
obtained. The kd values determined for tetrakis-NTA, however,
were probably too high due to the active exchange mechanism
as described above, which could account for the relative too
low ∆G°. Since the structure of tetrakis-NTA is much more
flexible compared to that of tris-NTA, a larger loss in entropy
upon complex formation could also be the reason for a lower
gain in ∆G°. However, not only flexibility of linkers49 but also
the different topologies have been shown to play a key role for
optimizing the free energy of multivalent complexes.47 For this
reason, substantial differences in ∆S° are expected for the
branched MCH bis-NTA and tetrakis-NTA, compared to that
of the cyclic tris-NTA.
contiguous stretch of histidines in case of redundant tags. Owing
to the degenerate interaction motif, however, the structures of
the complexes are probably not well defined because individual
histidines within the histidine tag can exchange coordination
positions, in particular in complexes with high redundancies.
Comparison of ∆G° and ∆H° of complex formation revealed
an increasing loss of entropy with increasing multivalency (i.e.,
declining gain in ∆G° relative to ∆H°). We ascribe this to the
high flexibility of both interaction partners, leading to a dramatic
loss of conformational freedom upon complex formation. Thus,
further improvement of binding affinities would require more
rigid scaffolds for the recognition unit. On the basis of these
insights, improving the affinity and selectivity of the interaction
with the MCH by tailoring tags with appropriately positioned
histidines appears promising. Other than mere affinity enhance-
ment, more ambitious differential binding could also be
envisioned. This interaction gives a synthetic control over the
design of the recognition unit and a recombinant combinatorial
control over the tags.
These high-affinity, yet switchable chemical recognition units
open a wide field of possible applications. In contrast to
conventional approaches based on protein-based recognition,8-11,14
both the MCH and the recognized oligohistidine peptide are
small, resulting in a highly localized attachment to the protein.
Spectroscopic probes of different nature can be coupled to the
MCH in a chemically and structural defined manner as already
demonstrated in this work. Other functional units such as lipids,
poly(ethylene glycols), and other polymers or nanoparticles
could be site-specifically coupled to proteins in a reversible
manner. Also for immobilization on surfaces, multivalent
chelators could provide distinct advantages.42 At the same time,
metal ions are incorporated into proteins, which could be used
as additional probes. Since the oligohistidine tag is the most
frequently applied affinity tag in heterologous protein expres-
sion, generic and widespread application of such conjugates is
guaranteed.
Conclusions
We have designed high-affinity recognition units for short
peptides by exploiting multivalent coordinative interaction with
the oligohistidine tag. A tremendous increase spanning 4 orders
of magnitude in affinity and complex stability was obtained for
the MCH compared to that of the conventional mono-NTA. At
the same time, fast reversibility of the interaction (switchability)
under mild conditions (imidazole, EDTA) was maintained,
which is an exceptional advantage of multivalent interactions.15
A prominent characteristic of most of the MCH/oligohistidine
complexes at 1:1 stoichiometry was an excess of coordination
sites either on the tag, or, in one case, on the chelator head.
Because of the already relatively high affinity of an individual
dihistidine/Ni:NTA interaction (∼10 µM), this feature, which
we called redundancy, has some important consequences. First,
complex stoichiometries should depend on the absolute con-
centrations: while clear 1:1 stoichiometries were observed at
low concentrations (<1 µM), other complex stoichiometries
cannot be excluded at high concentrations or at substantial
relative shortage in amount of the redundant partner. This feature
hampered precise assessment of equilibrium binding constants
by ITC, but opened new possibilities for actively exchanging
one ligand by another by docking the competitor through an
excessive binding site. Second, a substantial redundancy of
histidines was required to reach the theoretical, additive ∆H°,
indicating that steric constraints interfered with full coordination.
This observation suggests that the MCHs do not bind a
Acknowledgment. We thank Gerhard Spatz-Ku¨mbel for
excellent technical assistance in organic synthesis. We are
thankful to Christian Herrmann and Christof Hasse for help with
ITC measurements and evaluation and to Kai Schlepckow for
assistance with stopped-flow fluorescence measurements. Fluo-
rescence-labeled oligohistidine peptides were synthesized by
Thomas Andre´. We thank Michael L. Reese for critical reading
of the manuscript. This work was supported by the DFG within
the Emmy-Noether-program for young investigators (Pi 405/
1), within the research group “Chemisch-Biologische Hybrid-
verbindungen” (Pi 405/2, Ta 157/6), and by the BMBF within
the nanobiotechnology program (0312005A).
Supporting Information Available: Detailed description of
the synthesis, purification, and spectroscopic characterization
of the compounds, and a description of the fitting procedure
for reversible bimolecular interaction kinetics. This material is
available free of charge via the Internet at http://pubs.acs.org.
(48) Jencks, W. P. Proc. Natl. Acad. Sci. U.S.A. 1981, 78, 4046-4050.
(49) Glick, G. D.; Toogood, P. L.; Wiley, D. C.; Skehel, J. J.; Knowles, J. R.
J. Biol. Chem. 1991, 266, 23660-23669.
JA050690C
9
J. AM. CHEM. SOC. VOL. 127, NO. 29, 2005 10215