Second Messenger Pathways in Identified Neurons 
terminal domain that is highly enriched in pro- 
line, glutamine, serine, and threonine (PEST) res- 
idues. This sort of domain has been recognized as 
a determinant of proteins that turn over rapidly. 
But digestion of proteins does not normally re- 
quire the synthesis of new protein. Why then 
does the degradation of R subunits depend on 
new protein synthesis? 
Preliminary evidence suggests that the proteo- 
lytic mechanism responsible for the disappear- 
ance of R subunits under conditions of training 
depends upon ubiquitination, a complex degra- 
dative process involving enzyme-catalyzed cou- 
pling of a small protein, ubiquitin, to the protein 
to be destroyed. 
Since ubiquitin and the enzymes required for 
coupling are present in the neuron whether or 
not it has been stimulated for long-term facilita- 
tion, how is the degradation of R subunits en- 
hanced in stimulated neurons? Our working hy- 
pothesis is that the new protein synthesized as a 
consequence of long-term stimulation is a factor 
that targets R subunits for coupling with 
ubiquitin. 
A similar hypothesis has recently been ad- 
vanced for the periodic degradation of cyclin, a 
protein that regulates a special protein kinase 
crucially involved in cell division. In early frog 
embryos, cyclin is rapidly proteolyzed every 30 
minutes during mitosis. We find that R subunits 
of PKA in early Aplysia embryos also are degraded 
through ubiquitin-mediated proteolysis. In con- 
trast, the R subunit of mature muscle is quite 
stable. These observations that R subunits are 
readily down-regulated in growing cells and in 
nervous tissue provide suppon for the popular 
idea that both learning and development employ 
similar molecular mechanisms. 
392 
