A significant portion of these hypervariable regions are "orphaned" in that we do not know what particular smelly chemicals can bind to them. It would be great to get these receptors out of the orphanage because we could use these them in a test tube to detect e.g. explosives. The problem is that mammalian cells, for a variety of reasons, are not as efficient in such screens as cells traditionally involved in screening e.g. bacteria and yeast.
Enter engineered brewers yeast. By hijacking a rat G-protein signaling network, and coupling it to a genetic reporter, Radhika et al engineered a rat's nose in a (probably not but nonetheless potentially) delicious malty culture tube. Instead of changing the membrane potential as in an olfactory receptor cell, the yeast cells grow green in UV light. And as a proof of principle they de-orphaned a rat receptor, showing that it preferentially bound DNT, a toxic chemical associated with explosives.
Does this adopted receptor work the same way in rat cells? They didn't test. Can the system be used in a high-throughput screen for other toxins? Not yet. That's probably why it wasn't published in Science. I guess the big deal is that this is the first complex mammalian signaling network to be re-constituted in the way-evolutionarily distant yeast. In fact I imagine it is this evolutionary distance that affords the system so much power, as a similar system reconstituted in a more phylogenetically related cell might lead to unwanted molecular "cross talk" that could interfere with normal cellular physiology. Such an orthogonal signaling system affords another handy use for the lowly organism that booze consumers have grown to love.
For a professional perspective on this system check out, http://www.nature.com/nchembio/journal/v3/n6/full/nchembio0607-306.html
Note from the future: the yeast used in this might have been baker's yeast, S. cerevisiae. Too lazy to look it up though.