90 K. Mruk and W. R. Kobertz
Photoisomerizable bioreactive tethers offer several advantages over other op-
togenetic and optopharmacologic approaches. Compared to photocaged ligands,
azobenzene light-gated channels can be limited to specific cell types using tissue-
specific promoters. In addition, the azobenzene moiety can undergo multiple rounds
of photoswitching whereas the uncaging process is irreversible and can lead to ac-
cumulation of desensitized ligand-gated channels and receptors (Gorostiza and
Isacoff 2007 ; Kramer et al. 2005 , 2013 ). Although channelrhodopsins offer similar
cell-type specific expression, these channels have smaller single-channel currents
and cannot be actively turned off (Zhang et al. 2006 ). Moreover, sustained neu-
ronal excitation requires extended illumination with channelrhodopsins, which in
contrast occurs in the dark with LiGluR, limiting the potential for photodamage.
Therefore, azobenzene tethers provide superior spatiotemporal resolution compared
to the current complementary opto-technologies available. As the repertoire of azo-
benzene derivatives (Kienzler et al. 2013 ) and ligands (Levitz et al. 2013 ) continues
to expand, so do the possibilities for light control of ion channels in both in vitro
and in vivo systems.
3.2 Molecular Rulers
While photoswitchable tethers provide exquisite spatiotemporal control over chan-
nel function, small libraries of tethering agents have proven useful tools for probing
the structure and molecular movements of ion channels. Because the effective mo-
larity of the modifying reagent is directly dependent on tether length, these reagents
are often used as molecular calipers to sample the extracellular and intracellular
space around the ion channel (Fig. 4a). The first molecular calipers were devel-
oped to synthesize potent ligands that activate cyclic-nucleotide-gated channels that
have four nucleotide binding sites (Kramer and Karpen 1998 ). Instead of exposing
channels to random combinations of freely diffusing small molecules, the Karpen
lab generated libraries of polymer-linked ligand dimers (PLDs) in which ligands
were connected with variable length polyethylene glycol (PEG) tethers. Because
the effective concentration of the tethered ligands is directly proportional to tether
length, when the average length of the PLD matches the distance between binding
sites, the affinity increases leading to large changes in channel activation. Using this
approach, micromolar PLDs were found for olfactory and rod photoreceptor cyclic-
nucleotide-gated channels. Conversely, once the ideal length is established, synthet-
ic derivatives of cyclic nucleotides can be used to design compounds with various
pharmacological specificities and potencies. The authors also demonstrated that the
PLD approach works for the cytosolic protein, protein kinase G. Not surprisingly
the ideal distance for cyclic-nucleotide-gated channels (which contain binding sites
in individual subunits) was twice as long as those from protein kinase G (which
contains dual binding sites in a single subunit), suggesting this technique’s overall
utility for designing high-affinity bidentate ligands.