III. Additional Factors That Govern Complex StabilityA. STERICEFFECTS
Initially, ancillary ligands were envisioned as a way to impart
steric selectivity for a particular analyte of interest. By leaving
only two adjacent coordination sites available on the lanthanide,
for instance, it could be assumed that only bidentate ligands
would bind to the metal center and that tri- or tetradentate lig-
ands would be excluded from the lanthanide coordination sphere.
This design could be further extended by modification of the
ancillary ligand to generate a highly specific receptor site, with
hydrophobic or hydrophilic regions that would be complementary
only to the target analyte.
Our initial work with the dipicolinate system supports this
hypothesis. We found that a helper ligand leaving three linear
adjacent coordination sites on the terbium ion allowed for opti-
mal dipicolinate binding, while a ligand leaving three coordina-
tion sites in a triangular motif resulted in poor binding
(Fig. 12). This is consistent with dipicolinate chelation involving
oxygen donors from the two carboxyl units and the Pyr nitrogen
in a linear arrangement.
However, our work on various other systems has led us to reas-
sess our receptor designs. We have found that, in some cases, a
bidentate analyte will chelate more strongly to a lanthanide with
a hexadentate helper ligand (i.e., leaving three remaining coordi-
nation sites as opposed to two). Cat, the three CAs, and SU, all
bidentate analytes, bind more strongly to terbium complexes
involving a hexadentate ancillary ligand (DO2A and EDTA) as
opposed to a heptadentate one (DO3A). We also have discovered
that a cyclic helper ligand is preferred in some cases, while a lin-
ear (acyclic) ligand produces better results in others. A ligand
screen involving SA clearly indicated a greater binding affinity
for EDTA than cyclic DO2A (Fig. 13), and Job's method of contin-
uous variations ( 166 ) confirms a one-to-one binding stoichiome-
try of SA^2 – to Tb(EDTA)^3 – (Fig. 14). Further study, including
molecular modeling, may be necessary to elucidate the mecha-
nism that gives rise to such chelation preferences.
We conclude that the simple explanation based only on steric
interactions in the lanthanide coordination sphere is not suffi-
cient, and that there are more complex forces at work. We there-
fore recommend that ligand screens be performed to determine
the optimal ancillary ligand for a given target analyte when
designing lanthanide-based sensors.
30 MORGAN L. CABLEet al.