A pH-sensitive, colorful, lanthanide-chelating paramagnetic NMR probe

Article abstract of DOI:10.1021/ja307824e

Paramagnetic lanthanides ions are broadly used in NMR spectroscopy. The effects of unpaired electrons on NMR spectral parameters provide a powerful tool for the characterization of macromolecular structures and dynamics. Here, a new lanthanide-chelating NMR probe, Caged Lanthanide NMR Probe-7 (CLaNP-7), is presented. It can be attached to protein surfaces via two disulfide bridges, yielding a probe that is rigid relative to the protein backbone. CLaNP-7 extends the application range of available probes. It has a yellow color, which is helpful for sample preparation. Its effects are comparable to those of CLaNP-5, but its charge is two units lower (+1) than that of CLaNP-5 (+3), reducing the change in surface potential after probe attachment. It also has a different magnetic susceptibility tensor, so by using both tags, two sets of structural restraints can be obtained per engineered cysteine pair. Moreover, it was found that the orientation of the magnetic susceptibility tensor is pH dependent (pK_a ? 7) when a histidine residue is located in the neighborhood of the probe attachment site. The results show that the His imidazole group interacts with the CLaNP-7 tag. It is proposed that the histidine residue forms a hydrogen bond to a water/hydroxyl molecule that occupies the ninth coordination position on the lanthanide, thus breaking the two-fold symmetry of the CLaNP tag in a pH-dependent way.

Full text of DOI:10.1021/ja307824e

Article
pubs.acs.org/JACS
A pH-Sensitive, Colorful, Lanthanide-Chelating Paramagnetic NMR
Probe
Wei-Min Liu, Peter H. J. Keizers, Mathias A. S. Hass, Anneloes Blok, Monika Timmer, Alexi J. C. Sarris,
Mark Overhand, and Marcellus Ubbink*
Gorlaeus Laboratories, Leiden Institute of Chemistry, Leiden University, 2300 RA Leiden, The Netherlands
S
* Supporting Information
ABSTRACT: Paramagnetic lanthanides ions are broadly used in NMR spectroscopy. The
effects of unpaired electrons on NMR spectral parameters provide a powerful tool for the
characterization of macromolecular structures and dynamics. Here, a new lanthanide-
chelating NMR probe, Caged Lanthanide NMR Probe-7 (CLaNP-7), is presented. It can
be attached to protein surfaces via two disulfide bridges, yielding a probe that is rigid
relative to the protein backbone. CLaNP-7 extends the application range of available
probes. It has a yellow color, which is helpful for sample preparation. Its effects are
comparable to those of CLaNP-5, but its charge is two units lower (+1) than that of
CLaNP-5 (+3), reducing the change in surface potential after probe attachment. It also has
a different magnetic susceptibility tensor, so by using both tags, two sets of structural
restraints can be obtained per engineered cysteine pair. Moreover, it was found that the orientation of the magnetic susceptibility
tensor is pH dependent (pKa ≈ 7) when a histidine residue is located in the neighborhood of the probe attachment site. The
results show that the His imidazole group interacts with the CLaNP-7 tag. It is proposed that the histidine residue forms a
hydrogen bond to a water/hydroxyl molecule that occupies the ninth coordination position on the lanthanide, thus breaking the
two-fold symmetry of the CLaNP tag in a pH-dependent way.
possible configurations of the macrocycle rings (λλλλ and
δδδδ) and one of two orientations of the chelate arms (Δ and
Λ). The enantiomers Δλλλλ and Λδδδδ belong to the SA
structures, whereas the pair of Λλλλλ and Δδδδδ forms the
TSA conformation. In previous studies, we have developed
paramagnetic lanthanide probes based on cyclen.24−26 The
probes are attached to target proteins via two disulfide bridges,
which strongly reduces the mobility.24 In the most recent
version, two chelating pyridine-N-oxide arms force the probe to
exist into one diastereomer, the SA form, in the solution
state.25,27 This results in singular NMR resonances and one
magnetic susceptibility tensor. This probe, named Caged
Lanthanide NMR Probe-5 (CLaNP-5, Figure 1), showed the
largest PCSs and RDCs described so far.25 A drawback is the
total charge of the probe, which is 3+ after chelation with a
lanthanide ion. When CLaNP-5 is linked to a protein surface,
the extra charge changes the electrostatic potential, which may
affect molecular interactions with ligands and proteins. In the
current study, a new tag is described (CLaNP-7, Figure 1) that
has a reduced net charge while maintaining conformational
rigidity. Ln-CLaNP-7 complexes have different magnetic
susceptibilities than those of CLaNP-5 and its endogenous
color allows for easy detection of tagged protein. Furthermore,
we show that CLaNP-7 can interact with a nearby histidine
residue, causing a pH dependence of the magnetic susceptibility
tensor (χ-tensor). It was shown that two sets of distance
INTRODUCTION
■
In recent years, paramagnetic probes have been used broadly in
nuclear magnetic resonance (NMR) spectroscopy. The effects
of unpaired electrons on NMR spectral parameters, such as
pseudocontact shifts (PCSs), residual dipolar couplings
(RDCs), and nuclear relaxation enhancements, have been
recognized as powerful tools for the characterization of
macromolecular structures and their dynamics and interac-
tions.1−8 Among the paramagnetic effects, PCSs yield valuable
long-range distance and orientation information and RDCs are
distance-independent, providing information on the whole
protein, making them powerful restraints to refine protein
structures and determine protein orientations in complexes.9
To yield unambiguous restraints, the paramagnetic center
must be site-specifically and rigidly attached to the protein. For
Ca2+- or Mg2+-containing metalloproteins, lanthanide ions can
substitute the natural metal,10−12 but most proteins are devoid
of these metal binding sites. Therefore, several methods have
been developed to introduce artificial paramagnetic metals into
protein. For instance, metal binding peptide tags were designed
that can be engineered genetically at the N or C terminus or
into a loop region of a protein.13−17 Other strategies use
synthetic organic thio-functional tags, which can be introduced
via cysteine mutations.3,18,19
1,4,7,10-Tetraazacyclododecane-1,4,7,10-tetraacetic acid
(DOTA) is a typical chelator for lanthanide ions. The
coordination of DOTA to lanthanide ions results in two stable
diastereomers, a square antiprism (SA) and a twisted square
antiprism (TSA).20−23 Each diastereomer relates to one of two
Received: August 13, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja307824e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 1. Structures of Ln-CLaNP-524 and Ln-CLaNP-7.
a
Scheme 1
a
Reagents and conditions: (a) (i) BocOSu, CHCl3, rt, 36 h; (ii) methyl bromoacetate, K2CO3, rt, 12 h. (b) (i) TFA, DCM, rt, 4 h; (ii) 2-hydroxy-5-
nitrobenzyl bromide, K2CO3, ACN, 80 °C, 12 h. (c) NaOH, 1,4-dioxane, rt, 4 h. (d) Aminoethyl-MTS, NHS, EDC, DMF, rt, overnight. (e)
Ln(OAc)3, DMF, rt, 4 h, quant.
restraints can be obtained, simply by changing the pH. A model
for the pH-dependent interaction is proposed.
tetraazacyclododecane ring conformation.27 The yellow color
associated with p-nitrophenolate groups is convenient for the
purification of tagged proteins.
RESULTS
The synthesis of CLaNP-7 is shown in Scheme 1.
Commercially available cyclen was converted into 1 in 83%
yield over two steps, in which two opposing amines were
temporarily protected with tert-butoxycarbonyl (Boc) groups
and the other two amines subsequently functionalized by
reaction with methyl bromoacetate.29 Compound 1 was
deprotected by trifluoroacetic acid (TFA) and reacted with 2-
hydroxy-5-nitrobenzyl bromide at 80 °C to obtain 2.28,30
Removal of the methyl ester group in the presence of base in
1,4-dioxane as solvent yielded 3.31 Condensation of 3 with
excess (2-aminoethyl)methanethiosulfonate in presence of 1-
ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) and N-
hydroxysuccinimide (NHS) yielded CLaNP-7 in 40% yield.
The Yb-CLaNP-7, Tm-CLaNP-7, and Lu-CLaNP-7 complexes
were obtained by chelating with the corresponding lanthanide
acetate salts in N,N-dimethylformamide (DMF). After
chelation, the wavelength of the maximum absorbance shifted
from 310 to 390 nm, and the tag obtained a more intense
yellow color (Figure 2A).
■
Synthesis of Caged Lanthanide NMR Probe-7. In 2004,
Sherry’s group published a study on a magnetic resonance
imaging (MRI) contrast agent based on the cyclen scaffold and
carrying a p-nitrophenol ligating group.28 The crystal structure
of the MRI contrast agent indicated that the phenol is
deprotonated and a water molecule sits above the plane of the
oxygen atoms. In order to design a rigid cyclen based probe
with a reduction of net positive charge, the pyridine-N-oxide
arms used for CLaNP-5 were replaced with p-nitrophenol arms
(Figure 1). The p-nitro groups enhance the acidity of the
phenolic protons and upon binding the lanthanide, the phenol
groups will be deprotonated. As the lanthanide has a charge of
+3, the net charge of the entire complex will be +1, unless a
hydroxyl ion rather than a water molecule can coordinate, in
which case the overall charge is zero (see below). The two p-
nitrophenol groups provide C2 symmetry, and it was anticipated
that the two six-member ring chelating systems (Figure 1)
provide sufficient rigidity and thus lead to one preferred
B
dx.doi.org/10.1021/ja307824e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 2. Electronic absorption spectra of CLaNP-7. (A) The absorbance spectra expressed as extinction coefficients are plotted for CLaNP-7 and
Tm-CLaNP-7. (B) Lu-CLaNP-7 at different pH values.
Probe Attachment and Δχ-Tensors Calculation. The
Zn2+ form of 15N-enriched E51C/E54C pseudoazurin (Paz)
was reduced with dithiothreitol (DTT), washed, and directly
incubated with Ln-CLaNP-7 for 16 h at 4 °C. The tagged
protein was purified using a Superdex 75 column (GE
Healthcare). The mass of the resulting 15N-Paz Yb-CLaNP-7
(14 503 2 Da) agreed with the expected mass of 14 503 Da,
assuming 98% 15N enrichment. In the [15N,1H]-HSQC spectra,
there were no significant differences between untagged and Lu-
CLaNP-7 tagged spectra, except for a few residues close to the
attachment site, enabling the resonance assignments to be
made by comparison with previous spectra.32 Large shifts of
resonances were observed when the protein was tagged with
either Yb-CLaNP-7 or Tm-CLaNP-7 (Figure 3).
along with those of CLaNP-5 are reported in Table 1 and the
back-calculated PCSs are plotted versus the observed PCSs in
Figures 4A and S1, respectively. The values of Δχax and Δχrh of
CLaNP-7 are similar to those of CLaNP-5 for Tm, but not for
Yb. The Δχ-tensor of Yb-ClaNP-7 is much more rhombic. A
possible explanation could be the presence of a ninth ligand, a
water molecule or hydroxyl ion. This possibility is discussed in
more detail in the next section. Tm-CLaNP-7 causes significant
alignment of Paz at 600 MHz (14.1 T), allowing RDCs up to
20 Hz to be observed. The observed RDCs were used to
optimize the HN positions of Paz and the Δχ-tensor derived
from PCS was used to back-calculate the RDCs on the basis of
this structure, yielding a good agreement (Figure 4B). The large
RDCs indicate a low mobility of CLaNP-7 relative to the
protein backbone.
pH Dependence of CLaNP-7. In order to test CLaNP-7
on a second protein, yeast cytochrome c (Cyt c) N56C/L58C
was tagged with CLaNP-7 coordinated to Yb and Lu. In the
diamagnetic spectrum (Lu-CLaNP-7) only single peaks were
observed. However, at pH 7, the Cyt c tagged with Yb-CLaNP-
7 showed two sets of peaks for most residues. The relative
intensity and sometimes the sign as well changed with pH
(Figures 5 and S2). From the PCSs obtained at pH 6 and 8, the
Δχ-tensors were determined. A comparison of the PCSs shows
that they differ significantly (Figure S3). The sizes of the axial
and rhombic components are similar at both pH values and
somewhat larger than for Paz E51C/E54C (Table 1 and Figure
S4). The major difference between pH 6 and 8 is the
orientation of the tensors (Figure 6A). The structurally highly
similar tag Ln-CLaNP-5 shows a single set of resonances when
attached to the same double Cys mutant of Cyt c (Figure S5).33
Also for CLaNP-7 attached to Paz E51C/E54C no pH
dependence of the PCS was observed, so the observed effect
is specific for CLaNP-7 attached to Cyt c N56C/L58C.
Protonation of the phenol moiety of CLaNP-7 as an
explanation for the pH dependence was ruled out by the
electronic absorption spectrum. Protonation of the phenol
shifts the absorbance maximum from 390 to 310 nm.34 From
Figure 2A, it is clear that in free CLaNP-7 purified by HPLC
the phenol rings are protonated, but upon coordination to Tm
or Lu they are deprotonated at pH values between 3 and 11
(Figure 2B). In general, there is a water coordinated to the
DOTA-like lanthanide ions complex in solution state.35 The
factors that affect water coordination are mainly the charge
density of the lanthanide ion and the steric strain at the water
binding site.36−38 The substitution of the pyridine-N-oxide
rings in CLaNP-5 with p-nitrophenol rings in CLaNP-7 lowers
the charge of the coordinated oxygen of the ligand, which
Figure 3. Overlay of [15N,1H]-HSQC spectra of Paz E51C/E54C
attached to Lu-CLaNP-7 (blue) and Tm-CLaNP-7 (red). Several
PCSs are indicated with solid lines.
The difference in resonance frequency between the para-
magnetic and the diamagnetic samples was defined as the PCS.
The presence of large, single PCSs indicates the tag to be
attached rigidly and to exist in one dominant conformation,
which is expected to be the SA isomer due to the six-member
chelating system (but see Discussion section).27
For estimation of the anisotropic component of the magnetic
susceptibility tensor (Δχ-tensors), an initial metal position was
fixed according to a previously reported protocol.25 An initial
set of PCSs was used to determine the Δχ-tensor, with which
more PCSs were predicted. In an iterative way, the Δχ-tensor
and metal position were refined and additional resonances were
assigned. The Δχ-tensor values for Tm- and Yb-CLaNP-7,
C
dx.doi.org/10.1021/ja307824e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Table 1. PCSs-Based Δχ-Tensors of CLaNP-525 and CLaNP-7
b
b
c
protein
probe
CLaNP-5
Ln
Δχax
Δχrh
10
restraints
Q
Paz E51C/E54C
Tm
Yb
55.5 0.8
9.4 0.2
41.4 0.6
4.2 0.1
1
89
93
94
93
0.03
0.04
0.04
0.06
1.9 0.4
9.6 0.8
5.6 0.4
CLaNP-7
Tm
Yb
Cyt c N56/L58C
CLaNP-7 (pH 6)
CLaNP-7 (pH 8)
Yb
Yb
6.3 0.4
5.5 0.9
8.0 0.3
9.6 0.3
70
60
0.03
0.06
Cyt c N56C/L58C/H39A
CLaNP-7 (pH 6.3)
CLaNP-7 (pH 7.8)
Yb
Yb
5.5 0.2
5.5 0.1
6.5 0.1
6.1 0.1
70
62
0.04
0.04
a
b
The unit of axial and rhombic components Δχax and Δχrh is 10−32 m3. The error is calculated by a jackknife procedure, randomly removing 20% of
c
the data and repeating the Δχ-tensor fit 100 times. Q is defined in the Experimental Section (eq 2).
Figure 4. Experimentally observed amide proton PCSs and RDCs of Paz E51C/E54C Tm-CLaNP-7 plotted against the back-calculated PCSs (A, Q
= 0.04) and RDCs (B, Q = 0.09), both based on the PCS derived Δχ-tensor. The NMR spectra were recorded at 14.1 T (600 MHz). The solid line
represents a perfect match.
Figure 5. pH dependence of CLaNP-7. Detail of [15N,1H]-HSQC spectra of Cyt c N56C/L58C attached to Lu-CLaNP-7 (blue) and Yb-CLaNP-7
(red) at pH 6.0 (A), 7.0 (B), and 8.0 (C).
results in a larger space for a ninth coordination.27,28 Thus, it is
likely that a water molecule or hydroxyl ion can be coordinated
to the lanthanide.
pH values (Figure S6A) and the PCS and the Δχ-tensors are
the same at low and high pH (Table 1 and Figure S7). The
peaks only exhibit small shifts due to the normal pH
dependence of the protein and those shifts are identical in
paramagnetic and diamagnetic samples (Figure S6B). These
observations support the proposal that His 39 influences the
Δχ-tensor in a pH-dependent manner.
We wondered whether it was possible to introduce the pH
dependence of the tensor by positioning the probe close to a
His residue on a protein surface. For this purpose we returned
to Paz, which is a β-sheet protein. His 6 is located on a long β-
strand. In the neighboring strand two Cys residues were
engineered, mutant I34C/V36C. Using the published protocol
for modeling of the CLaNP lanthanide position,25 it was
predicted that the imidazole ring of His 6 would be sufficiently
close to form a H-bond to a possible H2O/OH at the ninth
In Cyt c N56C/L58C, the Cys residues are located in a short
β-strand and there is a histidine residue, His 39, in the
neighboring strand, close to the tag attachment site. Fitting of
the PCSs placed the lanthanide close to His 39, and its
imidazole ring can be rotated to be at hydrogen bond distance
of a water molecule at the ninth coordination position (Figure
6B). Therefore, we propose that the observed pH dependence
of the Δχ-tensor is caused by the imidazole of His 39.
In order to test this hypothesis, the His residue in Cyt c
N56C/L58C was mutated to alanine and this variant was
labeled with Ln-CLaNP-7. The pH dependence of the Δχ-
tensor is no longer observed in the absence of the imidazole
ring. The [15N,1H]-HSQC spectra shows single peaks at several
D
dx.doi.org/10.1021/ja307824e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
pH, the PCSs of both states also exhibit some pH dependence
(Figure S9).
In order to explain these observations, a four-state model is
invoked (Figure 8). The model hypothesizes that the two-fold
symmetry of the probe is broken by the interaction between the
imidazole of the nearby histidine and the proposed H2O/OH
coordinated to the Yb. Consequently, the two SA enantiomers
(Δλλλλ and Λδδδδ) no longer yield identical PCS and now
cause double peaks, because cyclen ring flips are normally
slow.40,41 Protonation/deprotonation of the imidazole shifts the
equilibrium between these two states. It also causes the pH
dependence of the PCSs of both states (Figure S9). The
intensity ratio (Ib/Ia) of the pairs of peaks can be described on
the basis of this model by eq 1:
Figure 6. PCS-based positions of the tensor relative to the Cyt c
structure with PDB entry code 1YCC.39 The Cys residues at positions
56 and 58 were modeled. (A) The principal axes of the Δχ-tensor of
Yb-CLaNP-7 are shown as red (pH 6) and blue sticks (pH 8). (B)
Model of CLaNP-7 attached to Cyt c N56C/L58C. The position of
the Yb (magenta) is derived from a fit to the PCS data. The red sphere
shows the H2O/OH at the ninth coordination position. The protein
main chain is shown in gray. The Cys (N56C/L58C) and the His 39
side chains are shown in yellow and CPK colors, respectively. The
carbon atoms from CLaNP-7 are shown in green, nitrogen atoms in
blue, oxygen atoms in red and sulfur atoms in yellow. The distance
between the Hε2 of His 39 and the ligating oxygen atom is 2.2 Å.
)
[1 + 10(pH−pK
[1 + 10(pH−pK
]
a2
Ib
Ia
= K
)
a1
]
(1)
where Ia and Ib are the peaks intensity of the two sets of peaks
and Ib is most intense at high pH, K is the equilibrium constant,
and pKa1 and pKa2 are the proton association constants. Fitting
the observed intensity ratio of the peaks for Cyt c N56C/L58C
to eq 1 yields the curve given in Figure 8B. The equilibrium
changes from K = 0.52 to K′ = 3.36, where K′ = K(Ka1/Ka2), so
state III becomes the dominant conformation when going from
high to low pH. For Paz I34C/V36C the effect is also clearly
present, although the shift in the equilibrium is smaller (Figure
8B).
coordination position (Figure S8). The HSQC spectra clearly
show two peaks for most resonances in this Paz variant labeled
with Yb-CLaNP-7 and the intensity of these peaks indeed varies
with pH, analogous to what was observed for Cyt c N56C/
L58C (Figure 7). Note that Paz E51C/E54C labeled with
Comparison with CLaNP-5. Tm-CLaNP-5 and Tm-
CLaNP-7 were attached to Paz E51C/E54C, and both caused
large PCSs and RDCs, but these shifts and couplings differ in
size and sometimes sign as well (Figure S10). The position of
the metal, as well as the size and orientation of the Δχ-tensor
were obtained by fitting to the PCS data. The PCS histograms
have an overall similar appearance for both tags because PCS is
dominated by the distance between the nucleus and the metal.
However, detailed inspection shows clear differences that reflect
the somewhat smaller size and changed orientation of the
CLaNP-7 Δχ-tensor. Also the positions of the Ln ions are
different by 3.4 Å (Figure 9C). The RDC histograms show
many more differences (Figure S10B), because RDC is solely
dependent on the H−N bond vector orientation in the Δχ-
tensor frame. The RDCs range from −15 to +20 Hz for Tm-
CLaNP-7 and from −20 to +30 Hz for Tm-CLaNP-5 at 14.1 T
(600 MHz). The observed PCSs and RDCs of CLaNP-7 were
plotted versus the observed PCSs and RDCs of CLaNP-5
(Figure 9) and showed poor agreement between CLaNP-5 and
CLaNP-7. It is concluded that the two probes cause sufficiently
different effects to yield complementary structural restraints
from a single attachment site.
Figure 7. Detail of [15N,1H]-HSQC spectra of Ln-CLaNP-7 tagged
15N-labeled Paz I34C/V36C. The spectra were recorded at pH 6.2
(black, Lu; red, Yb) and pH 7.9 (green, Lu; blue, Yb).
CLaNP-7 did not show pH dependence. These results indicate
that the pH dependence of the Δχ-tensor can be introduced
when the probe is positioned on a β-sheet, next to a histidine
residue. The pH dependence of CLaNP-7 may be used to
obtain two independent sets of PCSs or RDCs from a single
probe.
DISCUSSION
■
Proposed Mechanism for CLaNP-7 pH Dependence.
Based on the Δχ-tensor, the position of the Yb-CLaNP-7 linked
to Cyt c N56C/L58C is the same within error at pH 6 and 8.
However, the orientations of the tensors are different (Figure
6A). The two sets of peaks in the HSQC spectra of Yb-labeled
protein indicate that these orientations are a consequence of
two orientations/conformations of the probe that are in slow
exchange on the NMR chemical shift time scale. In contrast to
what is expected for a simple pH titration in which the two sets
of peaks would correspond to a protonated and deprotonated
form, here both sets of peaks remain (weakly) visible, even at
the highest and lowest pH values applied. Furthermore, not
only do the relative intensities of the pairs of peaks change with
CONCLUSION
■
CLaNP-7 is a new two-armed lanthanide chelating protein
probe that provides large PCSs and RDCs. CLaNP-7 and
CLaNP-5 have different magnetic susceptibility tensors,
providing a convenient way to obtain different paramagnetic
effects from the same target protein with a single set of
engineered Cys residues. In addition, the lower net charge of
CLaNP-7 will be beneficial in protein−protein interaction
studies. CLaNP-7 has a bright yellow color simplifying the
sample handling, and importantly, it is the first example of a
pH-sensitive paramagnetic probe offering a unique opportunity
to tune paramagnetic effects.
E
dx.doi.org/10.1021/ja307824e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 8. Proposed model to describe pH dependence of CLaNP-7. (A) The flip between two SA forms (Δλλλλ and Λδδδδ) are schematically
represented by a change in the cyclen ring (lines connecting the N atoms) and the ligand arms (round arrows). The values were obtained from a fit
to the Ib/Ia ratios of Cyt c N56C/L58C at multiple pH values (eq 1). (B) Intensity ratios as a function of pH. The averaged intensity ratios for pairs
of resonances in Cyt c N56C/L58C (red) and Paz I34C/V36C (blue) are plotted as a function of pH. The solid lines represent fits to the model
described in the text (eq 1). The intensity ratios are averaged from 13 residues for Cyt c and 10 residues for Paz and the error bars are the standard
deviation. The black line represents a standard deprotonation curve.
1364.7, 1157.3 cm−1. HR-MS: m/z 517.3225 [M+H]+, calcd
[C24H45N4O8] 517.3232.
1,7-Dimethlyene-(p-nitrophenol)-1,4,7,10-tetraazacyclodode-
cane-4,10-diacetic Acid (3). Compound 1 (400 mg, 0.75 mmol) was
EXPERIMENTAL SECTION
■
General. Ln(OAc)3, cyclen, 2-(aminoethyl)methanethiosulfonate
hydrobromide, 2-hydroxy-5-nitrobenzyl bromide, and all other
chemicals were used as purchased without further purification. Thin-
dissolved in a mixture of dichloromethane and TFA (4 mL, 1:3 v/v)
layer chromatography analysis was conducted on a DC-alufolien
and stirred for 4 h at room temperature. The reaction mixture was
instrument (Merck, Kieselgel60, F254) with detection by UV
concentrated under reduced pressure and coevaporated with toluene
absorption (254 nm). Flash chromatography was performed on
to remove the TFA. The crude mixture was dissolved in acetonitrile
Screening Devices silica gel 60 (0.04−0.063 mm). A Biocad Vision
HPLC instrument (PerSeptive Biosystems, Inc.) and an Akta Basic
FPLC instrument (GE Healthcare Inc.) were used for purifications.
Analytical, semipreparative, and preparative reversed-phase C18
columns were obtained from Phenomenex (Torrance, CA). Superdex
75, CM sepharose, and HiTrap SP columns were obtained from GE
Healthcare. NMR spectra were recorded on a Bruker AV-400 (400/
100 MHz) and Bruker Avance-III 600 (600/150 MHz) spectrometer.
A LCQ liquid chromatrography mass spectrometry system and a
Finnigan LTQ Orbitrap system were used for high-resolution mass
spectrometry and protein conjugation analysis. Fourier transform
infrared spectroscopy was performed on a Perkin-Elmer (Shelton, CT)
Paragon 1000 FTIR spectrometer. Melting points were obtained using
a SMP3 scientific melting apparatus (Stuart, Bibby Sterlin Ltd.)
Syntheses. 1,7-Di-Boc-1,4,7,10-tetraazacyclododecane-4,10-di-
methyl Acetate (1). BocOSu (2.49 g, 11.6 mmol) was dissolved in
dry CHCl3 (20 mL) and was added under argon to a cyclen (1 g, 5.8
mmol) solution in CHCl3 (40 mL). The reaction mixture was stirred
at room temperature for 36 h. To the reaction mixture 3 M NaOH (30
mL) was added and stirring was continued for 30 min. Brine (30 mL)
was added and the solution was extracted with dichloromethane (3 ×
20 mL). The organic fractions were combined, dried over MgSO4 and
concentrated in vacuo. The remaining colorless oil was dissolved in
acetonitrile (30 mL). K2CO3 (1.94 g, 14.5 mmol) and methyl 2-
bromoacetate (2.22 g, 14.5 mmol) were added into the solution and
stirred for 12 h at room temperature. The reaction mixture was
concentrated, dissolved in ethyl acetate and washed with brine. The
organic layer was dried (MgSO4), filtered and concentrated. The
resulting residue was purified by column chromatograghy to yield
compound 1 as a white amorphous solid (2.48 g, 83% yield), mp =
(10 mL). 2-Hydroxy-5-nitrobenzyl bromide (435 mg, 1.88 mmol) and
K2CO3 (260 mg, 1.88 mmol) were added into the solution, which was
stirred for 12 h at 80 °C. The reaction mixture was filtered and the
solid was washed with acetonitrile (30 mL). The filtrate was
concentrated in vacuo to give a yellow oil. The crude compound was
dissolved in a solution mixture which contained 1,4-dioxane (final
concentration was 15 mM) and 3 M NaOH (final concentration was
0.4 M). The solution was stirred at room temperature for 4 h, then
concentrated under reduced pressure and purified by HPLC (0.1%
TFA and a 10−50% acetonitrile gradient on C18 preparative column).
1H NMR (400 MHz CD3OD): δ = 3.19 (s, 8H), 3.33 (s, 4H), 3.44 (s,
8H), 4.64 (s, 4H), 7.12 (d, 2H), 8.30 (d, 2H), 8.54 (s, 2H). 13C NMR
(100 MHz CD3OD): δ = 174.71, 164.29, 142.31, 130.87, 129.33,
117.59, 117.39, 54.15, 51.57, 49.87. FTIR: 3084.3, 1728.1, 1667.9,
1593.1, 1496.0, 1338.9, 1286.1, 1185.8, 1133.0, 1084.8 cm−1. HR-MS:
m/z 591.2409 [M+H]+, calcd [C26H35N6O10] 591.2409.
CLaNP-7. Compound 3 (296 mg, 0.5 mmol) was dissolved in DMF
(10 mL) and treated with 2-(aminoethyl)methanethiosulfonate
hydrobromide (283.4 mg, 1.2 mmol), NHS (345.3 mg, 3 mmol),
and EDC (574.5 mg, 3 mmol). The reaction mixture was stirred
overnight at room temperature. The reaction mixture was concen-
trated under reduced pressure and purified by HPLC (0.1% TFA and a
20−45% acetonitrile gradient on C18 preparative column) yielding
1
CLaNP-7 (40%) as a yellow oil. H NMR (600 MHz, DMSO-d6, 323
K): δ = 3.09 (s, 8H), 3.21 (s, 4H), 3.27 (s, 8H), 3.31 (t, 4H), 3.47−
3.53 (br, m, 10H), 7.12 (d, 2H), 8.21 (q, 2H), 8.47 (s, 2H), 8.51 (s,
2H). 13C NMR (150 MHz, DMSO-d6, 323 K): δ = 34.87, 38.21, 48.68,
49.80, 50.17, 51.52, 53.92, 116.40, 127.11, 129.22, 139.12, 164.06,
169.98. FTIR: 3465.8, 1649.8, 1138.9, 1198.9, 1134.5, 1100.8 cm−1.
HR-MS: m/z 865.2358 [M+H]+, calcd [C32H49N8O12S4] 865.2347.
Lu-CLaNP-7. Lu(OAc)3 (2.37 mg, 5.38 μmol) and CLaNP-7 (4.19
mg, 4.84 μmol) were separately dissolved in DMF (50 μL). Then, the
CLaNP-7 solution was added into Lu(OAc)3 solution and stirred for 4
h at room temperature. Without further purification, Lu-CLaNP-7 was
1
82−83 °C, Rf = 0.3 (5% MeOH in DCM). H NMR (400 MHz
CDCl3): δ = 1.27 (s, 18H), 2.69 (br, 8H), 3.21 (br, 8H), 3.28 (s, 4H),
3.51 (s, 6H). 13C NMR (100 MHz CDCl3): δ = 28.14, 46.21, 50.98,
54.36, 54.54, 79.05, 155.51, 171.39. FTIR: 2975.8, 1739.6, 1683.9,
F
dx.doi.org/10.1021/ja307824e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
primers, respectively. The expression and purification of the 15N
enriched variant was performed using the same conditions as described
above. The final yield obtained for Paz I34C/V36C was 2.4 mg/L of
culture. 15N enriched yeast Cyt c N56C/L58C was produced and
purified by the published method.33 To create mutant Cyt c H39A/
N56C/L58C, the forward and backward primers 5′-CTTGCATGG-
TATCTTTGGCAGAGCCTCTGGTCAAGCTGAAGGGTATTCG-
3′ and its complement were used on the template encoding Cyt c
N56C/L58C. Expression and purification was identical to that of
N56C/L58C. The final yield was 2.8 mg/L of culture.
Paramagnetic Probe Attachment. To attach Ln-CLaNP7 to Paz
and Cyt c, protein sample (1 mL, 150−300 μM) was treated with
DTT (final concentration 5 mM) at 0 °C for 1 h to remove possible
dimers. The reaction mixture was loaded on a PD-10 column (GE
Healthcare) pre-equilibrated with labeling buffer (20 mM sodium
phosphate, 150 mM NaCl, pH 7.0) to remove DTT. To avoid any
reoxidation by air, the buffer was degassed and the PD-10 column kept
under an argon atmosphere. To the eluted protein five equivalents Ln-
CLaNP-7 were added. The solution was stirred overnight at 4 °C. The
probe attached Paz sample was concentrated to 500 μL and purified
over a Superdex 75 gel filtration column. In the case of Cyt c, the
protein was purified on a HiTrap-SP column. The yield of labeling,
estimated from the intensity of diamagnetic peaks in the [15N,1H]-
HSQC spectra of samples with paramagnetic tags, was more than 90%.
NMR Spectroscopy. The NMR samples of Paz Ln-CLaNP-7
(100−200 μM) were prepared in 20 mM sodium phosphate, 150 mM
NaCl buffer, and 6% (v/v) D2O. All Cyt c samples (100−200 μM)
contained 20 mM sodium phosphate buffer, 6% (v/v) D2O and 1
equiv of ascorbic acid under an argon atmosphere to keep the Cyt c in
the reduced state. The pH of these samples was set with small aliquots
of 0.1 M HCl or 0.1 M NaOH and was checked before each titration
point. All [15N,1H]-HSQC and IPAP spectra43 were recorded at 298 K
on a Bruker Avance III 600 MHz spectrometer. Data were processed
with NMRPipe44 and analyzed with CCPNMR Analysis version
2.1.45Assignments of the resonances were based on previous
work.32,46,47
PCS and RDC Analysis. PCS were defined as the chemical shift
difference for a resonance in the paramagnetic and diamagnetic
sample. Positioning of the metal and optimization of the Δχ-tensor
were preformed as described for CLaNP-5,25 by using XPLOR-NIH
version 2.9.948 and Pararestraints.49 The structures of Paz and Cyt c
were taken from PDB entries 1PY0 and 1YCC and hydrogens were
added.32,39 The variation in the Ln positions were calculated by
randomly removing 20% PCS data and repeating the Δχ-tensor fit 100
times. For RDC analysis, the HN positions were optimized on the
basis of the experimental RDCs. The optimized structure was used for
back-calculation of the RDCs on the basis of the PCS-derived Δχ-
tensor. The Q factor provides a normalized measure for the agreement
between a set of observed and calculated experiment data:50
Figure 9. Comparison of CLaNP-5 and CLaNP-7. Experimentally
observed PCSs and RDCs from Tm-CLaNP-7 plotted against the
experimentally observed PCSs (A, Q = 0.2) and RDCs (B, Q = 0.5)
from Tm-CLaNP-5 for Paz E51C/E54C. The solid line represents a
perfect correlation. (C) Experimentally determined position and
orientation of the Tm Δχ-tensor relative to the Paz (PDB entry
1PY0)32 structure for CLaNP-5 (red sticks) and for CLaNP-7 (blue
sticks). The Cys side chains are shown in CPK colors.
2
Q =
(Oobs − Ocalc)2/ (Oobs + Ocalc
)
∑
∑
i
i
i
i
(2)
i
i
where Oiobs and Oicalcd are the observed and calculated PCSs or RDCs.
ASSOCIATED CONTENT
■
used to label protein samples. The other lanthanide ions Yb, and Tm
were chelated to CLaNP-7 following the same procedure. HR-MS: m/
z 1036.1513 [M−2H]+, calcd [C32H46N8O12S4Lu] 1037.1520.
S
* Supporting Information
[15N,1H]-HSQC spectra of CLaNP-7- or CLaNP-5-tagged
Cyt c; experimentally observed PCSs plotted against the
backcalculated PCSs for Cyt c and Paz; model of CLaNP-7
linked to Paz I34C/V36C mutant; details of PCSs and RDCs
data. This material is available free of charge via the Internet at
http://pubs.acs.org.
Protein Production and Purification. The production and
purification of the 15N enriched Alcaligenes faecalis pseudoazurin (Paz)
double cysteine mutant E51C/E54C was performed as described
before,32 with small modifications. Instead of Escherichia coli BL21
(DE3), E. coli BL21 (PlysS) was used to produce Paz E51C/E54C and
Tris-HCl buffer (20 mM Tris-HCl, pH 7.0) was used for ion exchange
chromotagraphy on the CM column. Double cysteine mutations
I34C/V36C were prepared by the QuikChange method using the
expression plasmid of wt Paz as a template.42 The oligonucleotides 5′-
GGCGACACGGTCACCTTTTGTCCGTGCGACAAAGGACA-
TAATG-3′ and its complement were used as the forward and reverse
AUTHOR INFORMATION
■
Corresponding Author
m.ubbink@chem.leidenuniv.nl
G
dx.doi.org/10.1021/ja307824e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Ogrin, D.; Tircso,
́
G.; Zhang, S.; Zhao, P.; Sherry, A. D. J. Am. Chem.
Notes
Soc. 2004, 126, 9248−9256.
The authors declare no competing financial interest.
(29) De Leon-Rodríguez, L. M.; Kovacs, Z.; Esqueda-Oliva, A. C.;
́
Miranda-Olvera, A. D. Tetrahedron Lett. 2006, 47, 6937−6940.
(30) Dapporto, P.; Fusi, V.; Micheloni, M.; Palma, P.; Paoli, P.;
Pontellini, R. Inorg. Chim. Acta 1998, 275−276, 168−174.
(31) Li, C.; Winnard, P., Jr; Bhujwalla, Z. M. Tetrahedron Lett. 2009,
50, 2929−2931.
ACKNOWLEDGMENTS
■
Financial support was received from the Netherlands
Organisation for Scientific Research, grants 700.58.405
(P.H.J.K.), 700.10.407 (M.A.S.H.), and 700.58.441 (W.-M.L.,
M.T., and M.U.)
̂
(32) Prudencio, M.; Rohovec, J.; Peters, J. A.; Tocheva, E.;
Boulanger, M. J.; Murphy, M. E. P.; Hupkes, H.-J.; Kosters, W.;
Impagliazzo, A.; Ubbink, M. Chem.Eur. J. 2004, 10, 3252−3260.
(33) Xu, X.; Keizers, P. H. J.; Reinle, W.; Hannemann, F.; Bernhardt,
R.; Ubbink, M. J. Biomol. NMR 2009, 43, 247−254.
(34) Ni, W.; Fang, H.; Springsteen, G.; Wang, B. J. Org. Chem. 2004,
69, 1999−2007.
REFERENCES
■
(1) Bertini, I.; Luchinat, C.; Parigi, G.; Pierattelli, R. ChemBioChem
2005, 6, 1536−1549.
(2) Clore, G. M. Protein Sci. 2011, 20, 229−246.
(3) Keizers, P. H. J.; Ubbink, M. Prog. NMR Spectrosc. 2011, 58, 88−
96.
(4) Otting, G. Annu. Rev. Biophys. 2010, 39, 387−405.
(5) Pintacuda, G.; John, M.; Su, X.-C.; Otting, G. Acc. Chem. Res.
2007, 40, 206−212.
(6) Su, Y.; Hu, F.; Hong, M. J. Am. Chem. Soc. 2012, 134, 8693−
8702.
̌
(35) Polas
́
e
̌
k, M.; Sedinova,
́
M.; Kotek, J.; Vander Elst, L.; Muller, R.
N.; Hermann, P.; Lukes,
̌
I. Inorg. Chem. 2009, 48, 455−465..
(36) Green, K. N.; Viswanathan, S.; Rojas-Quijano, F. A.; Kovacs, Z.;
Sherry, A. D. Inorg. Chem. 2011, 50, 1648−1655.
(37) Parker, D.; Puschmann, H.; Batsanov, A. S.; Senanayake, K.
Inorg. Chem. 2003, 42, 8646−8651.
(38) Zhang, S.; Jiang, X.; Sherry, A. D. Helv. Chim. Acta 2005, 88,
923−935.
(7) Bernini, A.; Spiga, O.; Venditti, V.; Prischi, F.; Botta, M.; Croce,
G.; Tong, A. P.-L.; Wong, W.-T.; Niccolai, N. J. Inorg. Biochem. 2012,
112, 25−31.
(39) Louie, G. V.; Brayer, G. D. J. Mol. Biol. 1990, 214, 527−555.
(40) Aime, S.; Barge, A.; Botta, M.; Fasano, M.; Danilo Ayala, J.;
Bombieri, G. Inorg. Chim. Acta 1996, 246, 423−429.
(41) Jacques, V.; Desreux, J. F. Inorg. Chem. 1994, 33, 4048−4053.
(42) Ho, S. N.; Hunt, H. D.; Horton, R. M.; Pullen, J. K.; Pease, L. R.
Gene 1989, 77, 51−59.
(43) Tjandra, N.; Omichinski, J. G.; Gronenborn, A. M.; Clore, G.
M.; Bax, A. Nat. Struct. Mol. Biol. 1997, 4, 732−738.
(44) Delaglio, F.; Grzesiek, S.; Vuister, G. W.; Zhu, G.; Pfeifer, J.;
Bax, A. J. Biomol. NMR 1995, 6, 277−293.
(45) Vranken, W. F.; Boucher, W.; Stevens, T. J.; Fogh, R. H.; Pajon,
A.; Llinas, M.; Ulrich, E. L.; Markley, J. L.; Ionides, J.; Laue, E. D.
Proteins 2005, 59, 687−696.
(8) Kobashigawa, Y.; Saio, T.; Ushio, M.; Sekiguchi, M.; Yokochi, M.;
Ogura, K.; Inagaki, F. J. Biomol. NMR 2012, 53, 53−63.
(9) Otting, G. J. Biomol. NMR 2008, 42, 1−9.
(10) Bertini, I.; Gupta, Y. K.; Luchinat, C.; Parigi, G.; Peana, M.;
Sgheri, L.; Yuan, J. J. Am. Chem. Soc. 2007, 129, 12786−12794.
(11) Biekofsky, R. R.; Muskett, F. W.; Schmidt, J. M.; Martin, S. R.;
Browne, J. P.; Bayley, P. M.; Feeney, J. FEBS Lett. 1999, 460, 519−
526.
(12) Gay, G. L.; Lindhout, D. A.; Sykes, B. D. Protein Sci. 2004, 13,
640−651.
(13) Barthelmes, K.; Reynolds, A. M.; Peisach, E.; Jonker, H. R. A.;
DeNunzio, N. J.; Allen, K. N.; Imperiali, B.; Schwalbe, H. J. Am. Chem.
Soc. 2011, 133, 808−819.
(46) Xu, X.; Reinle, W.; Hannemann, F.; Konarev, P. V.; Svergun, D.
I.; Bernhardt, R.; Ubbink, M. J. Am. Chem. Soc. 2008, 130, 6395−6403.
(47) Worrall, J. A. R.; Kolczak, U.; Canters, G. W.; Ubbink, M.
Biochemistry 2001, 40, 7069−7076.
(14) Gaponenko, V.; Dvoretsky, A.; Walsby, C.; Hoffman, B. M.;
Rosevear, P. R. Biochemistry 2000, 39, 15217−15224.
(15) Ma, C.; Opella, S. J. J. Magn. Reson. 2000, 146, 381−384.
(16) Barb, A. W.; Ho, T. G.; Flanagan-Steet, H.; Prestegard, J. H.
Protein Sci. 2012, 21, 1456−1466.
(48) Schwieters, C. D.; Kuszewski, J. J.; Tjandra, N.; Clore, M. G. J.
Magn. Reson. 2003, 160, 65−73.
(49) Banci, L.; Bertini, I.; Cavallaro, G.; Giachetti, A.; Luchinat, C.;
Parigi, G. J. Biomol. NMR 2004, 28, 249−261.
(17) Saio, T.; Ogura, K.; Shimizu, K.; Yokochi, M.; Burke, T.;
Inagaki, F. J. Biomol. NMR 2011, 51, 395−408.
(18) Swarbrick, J. D.; Ung, P.; Su, X.-C.; Maleckis, A.; Chhabra, S.;
Huber, T.; Otting, G.; Graham, B. Chem. Commun. 2011, 47, 7368−
7370.
(50) Bashir, Q.; Volkov, A. N.; Ullmann, G. M.; Ubbink, M. J. Am.
Chem. Soc. 2010, 132, 241−247.
̂
̀
(19) HaIussinger, D.; Huang, J.-r.; Grzesiek, S. J. Am. Chem. Soc.
2009, 131, 14761−14767.
(20) Aime, S.; Botta, M.; Ermondi, G. Inorg. Chem. 1992, 31, 4291−
4299.
(21) Aime, S.; Botta, M.; Fasano, M.; Marques, M. P. M.; Geraldes,
C. F. G. C.; Pubanz, D.; Merbach, A. E. Inorg. Chem. 1997, 36, 2059−
2068.
(22) Cosentino, U.; Villa, A.; Pitea, D.; Moro, G.; Barone, V.;
Maiocchi, A. J. Am. Chem. Soc. 2002, 124, 4901−4909.
(23) Ranganathan, R. S.; Raju, N.; Fan, H.; Zhang, X.; Tweedle, M.
F.; Desreux, J. F.; Jacques, V. Inorg. Chem. 2002, 41, 6856−6866.
(24) Keizers, P. H. J.; Desreux, J. F.; Overhand, M.; Ubbink, M. J.
Am. Chem. Soc. 2007, 129, 9292−9293.
(25) Keizers, P. H. J.; Saragliadis, A.; Hiruma, Y.; Overhand, M.;
Ubbink, M. J. Am. Chem. Soc. 2008, 130, 14802−14812.
(26) Vlasie, M. D.; Comuzzi, C.; van den Nieuwendijk, A. M. C. H.;
Pruden
1715−1723.
(27) Polas
K.; Lukes,
̂
cio, M.; Overhand, M.; Ubbink, M. Chem.Eur. J. 2007, 13,
́
e
̌
k, M.; Kotek, J.; Hermann, P.; Císarova, I.; Binnemans,
̌
́
̌
I. Inorg. Chem. 2009, 48, 466−475..
(28) Woods, M.; Kiefer, G. E.; Bott, S.; Castillo-Muzquiz, A.;
Eshelbrenner, C.; Michaudet, L.; McMillan, K.; Mudigunda, S. D. K.;
H
dx.doi.org/10.1021/ja307824e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX