Biologically inert caged GABA

Since the invention of CNB-caged glutamate and GABA by George Hess in 1994, many others labs have introduced other caged versions using different photochemical protecting groups. See my Beilstein J. Org. Chem. 2013, 9, 64–73 review for a summary of work up to 2013. In 2000 Molnar reported (Eur J Pharmacol 2000 vol. 391 (3) pp. 255-62) that CNB-GABA blocked GABA-A receptors in neurons (IC-50 = 0.028 mM).

When Hess made the CNB probes, they were really concerned to be able to uncage very quickly, as before their contribution, one had to use rapid flow techniques to deliver agonists quickly. Hess was a biochemist and so did not have all the concerns a neuroscientist might have. Molnar’s contribution was important, but largely ignored.

After the development of “NI-GABA” in 2001 by Corrie (J Neursci Meth 2001 vol. 112 pp. 29-42), it was obvious that caged GABAs with a normal protecting group were antagonists (Note Corrie does not cite Molnar). At this time Ogden said that MNI-Glu was not an antagonist. Later, Kasai discovered that MNI-Glu was a strong antagonist towards GABA-A at concentrations required for 2P uncaging. It never occurred to us this might happen.

From that time two groups have worked on the development of “biologically inert” (= low GABA-A antagonism) caged neurotransmitters. Both groups took the same approach. The idea was that adding “more negative charge” to GABA (or Glu) would stop antagonism. Thus, appending two phosphates was the most obviously efficient way to “bulk up” anionic density. In 2009 Ogden and Corrie revealed that this did not really work very well (J Neurosci Meth 2009 vol. 181 (2) pp. 159-69).

Ogden measured mIPSCs from interneurons in cultured neurons. Unfortunately 1-2 mM significantly reduced the amplitude of spontaneous events (the frequency was not reported). 2 mM was the highest concentration reported (5-10 mM would be needed for 2-photon uncaging) but the amplitude reduction was less than 1 mM NI-GABA, suggesting this was not baseline for their protocol. Their estimate for the IC-50 was 0.5 mM, similar to CDNI-GABA (but see below). Our paper using the latter was in preparation then, and we had made bisphosphate versions of MNI-GABA and CDNI-GABA independently. But Ogden’s result meant we gave up on this, even though our synthesis was better.
In 2009 Yuste and Etchenique reported RuBi-Glu was much better than MNI-Glu for the release of Glu under conditions with mild antagonism (Front. Neural Circuits 2009 vol. 3 pp. 2).

Previously they had revealed that RuBi-GABA (Front. Neural Circuits 2008 vol. 2 pp. 2) could be used for 1P uncaging at concentrations that were biologically inert (0.02 mM had no effect on mIPSCs). Note we have determined the IC-50 for RuBi-GABA is 0.004 mM. So there is some contradiction here. Recently, Higley has determined the CDNI-GABA IC-50 to be 0.11 mM (stronger than Matsuzaki in 2010) and CDNI-Glu to be 0.24 mM. Thus, using mIPSCs and evoked IPSCs give rather different values.

Certainly RuBi-GABA is very effective for 1P uncaging with blue light. Yuste says it is quite toxic at concentrations above 0.02 mM.

In 2012 Goeldner made a nitrophenylpropyl-caged GABA that worked for 2P uncaging when it was applied at 1 mM, but this probe blocked minis at 0.1 mM in culture (Angew. Chem. Int. Ed. 2012, 51, 1840 –1843 ).

So even though there are several caged GABA probes that work quite well, there is no probe that works well for 2PE at concentrations that are inert, until now.

In December 2016 we published a new approach to the reduction of GABA-A antagonism, which we called “cloaked caged compounds”. We conjugate a large dendrimer to the caged compound, this allows us to apply 0.2 mM without any antagonism.

We estimate the IC-50 is 0.9 mM, much better than the PEG version of DEAC450-GABA (0.011 mM).

Obviously we are testing this approach on other caging chromophores.