ABSTRACT
Our first observation of a chemical exchange saturation transfer
(CEST) signal from a paramagnetic lanthanide complex came when
one of my chemistry PhD students, Kuangcong Wu, collected a high-
resolution nuclear magnetic resonance (NMR) spectrum of the Eu3+
complex, EuDOTA-(gly-OEt)3+4 , in a mixture of water/acetonitrile (the ligand DOTA-(gly-OEt)4 is a tetraamide derivative of DOTA).
Much to our surprise, the 1H NMR spectrum of this complex
showed eight hyperfine shifted ligand proton resonances over the
chemical shift range covering+24 to−18 ppm (as expected) plus an unexpected resonance near +50 ppm with an integrated intensity of two protons. No one at the time would have considered the
possibility that this extra resonance could reflect a water molecule
coordinated directly to the Eu3+ in the complex because water exchange in all known lanthanide complexes was thought to be
much too fast for a separate resonance to appear. A water resonance
at +50 ppm in a 1H 500 MHz NMR spectrum would dictate the rate of water exchange to be <<1.6 × 105 s−1, about 10-fold
slower than any other known lanthanide complex at that time. Yet,
further investigations proved that the resonance at +50 ppm did indeed reflect a single Eu3+-bound water molecule. One experiment Kuangcong did in an effort to assign the peak at +50 ppm was to apply a long frequency-selective RF pulse on the shifted water
resonance to see whether it was spin-coupled to another resonance.
Saturation of this peak did not alter the intensities of the other ligand
peaks but did result in substantial reduction in the intensity of the
solvent water resonance; hence, this was the first CEST experiment
performed in Dallas! The details of this experiment were described
in our first paraCEST paper in 2001 [1]. We quickly realized the
beauty of having an exchanging proton peak +50 ppm away from the water signal was that it could be saturated with negligible off-
resonance saturation of water so that any decrease in water signal
could be directly ascribed to CEST. This suggested that one should be
able to create responsive paraCEST agents and detect them easily by
simple “off” minus “on” image intensity differences. We immediately
set out to measure Z-spectra for each of the other paramagnetic
lanthanide ion complexes with DOTA-(gly-OEt)4 hoping that they
would all display equally slow water exchange kinetics favorable
for CEST. This would provide us with a family of paraCEST agents
with water exchange peaks suitable for CEST activation over an
incredibly wide frequency range (± 600-700 ppm). Although we succeeded in detecting a water exchange CEST peak for the entire
series of lanthanide complexes (except the Yb3+ complex, which as we learned later does not have an inner-sphere water molecule),
the lanthanide complexes that produce the largest hyperfine shifts
(Tb, Dy, Ho, Tm) all displayed extraordinarily broad water exchange
peaks and required extremely high-power B1 pulses for detection
[2]. This was the first indication that water exchange in these
complexes must be highly variable and in some cases quite fast.
A later more complete study of the water exchange kinetics of
the entire lanthanide series of complexes revealed that the Eu3+
complex surprisingly displayed the slowest rate of water molecule
exchange of all the lanthanide complexes [3]. The origins of this
effect have not been fully proven, but based on the elegant work of
Merbach et al. [4-6] on thermodynamicmeasures ofwater exchange,
it is reasonable to conclude that water exchange is slowest for Eu3+
because this ion is near the optimum size for the DOTA-(gly-OEt)4 macrocyclic cavity and this size match forces the mechanism of
water exchange to switch from an associative mechanism for the
larger lanthanide ion (La3+, Ce3+, Pr3+, and Nd3+) complexes to a dissociative mechanism at Eu3+. This switch in water exchange mechanism could, of course, differ for other types of CEST ligands
having different metal ion binding cavity sizes, but this has yet to be
explored in sufficient detail for other paraCEST systems. In general,
it appears to hold that water exchange is slowest for all Eu-based
CEST complexes reported so far (compared to other lanthanide
complexes with the same ligand), while exchange appears to be
generally faster for the smaller lanthanide ion complexes (Tb3+, Dy3+, Ho3+, Er3+, Tm3+, and Yb3+).