So this was a very exciting period. So we wanted to identify some of the molecular mechanisms underlying this. So we began to test this idea. So - we knew that the cyclic AMP dependent protein kinase could activate a particular transcription factor called CREB - cyclic AMP response element-binding protein. And how that happened was that the cyclic AMP dependent protein kinase moved into the nucleus to activate CREB. So Roger Tsien had developed imaging techniques and he relabelled the cyclic AMP dependent protein kinase, and we actually saw it move into the nucleus with repeated serotonin applications.

And then, first Pramod Dash and then Dusan Bartsch showed that CREB was critically involved. Pramod Dash just showed that if you inject a blocker of CREB it blocks long-term memory. Dusan Bartsch actually cloned CREB and showed that if you inject the phosphorylated form of CREB you don't need anything; you can get long-term facilitation. Doesn't last very long, but you can turn it on. Moreover he made an amazing discovery. He found there were two CREBs: CREB1 and CREB2. CREB1 activates transcription, and CREB2 inhibits CREB1. That was fascinating. And he realized that there was a machinery for turning on the long-term process. And that machinery became even more interesting about five or six years ago when Priya Rajasethupathy, a fantastic MD-PhD student was in my lab, and she found that there are small RNAs that regulate CREB1 and CREB2. There's one, a micro RNA, that normally inhibits CREB1, and serotonin releases that inhibition, turning on CREB1. But in addition there's a piRNA which no one realized exists in a nervous system, a whole bunch of them. And one of them inhibits CREB2 and gets turned on by serotonin. So serotonin relieves the inhibition in CREB1, turns on the inhibition of CREB2, so CREB1 can be on for longer period of time to turn on the long-term process. Really quite beautiful. So with time we've got a very good understanding of how that transcriptional cascade works, and the growth of new synaptic connections.

And that has turned out to be a very general mechanism. In almost every learning process, simple or complicated - Aplysia, hippocampus - that one looks at, long-term memory involves CREB. May not be by itself, but acting with other, you know, colleagues, but it's a critical factor in it. But that generality was reassuring, but it raised an important problem in the cell biology of neurons.