To explore the role of protein kinase A (PKA) in the late phase of long-term potentiation (L-LTP) and memory, we generated transgenic mice that express R(AB), an inhibitory form of the regulatory subunit of PKA, only in the hippocampus and other forebrain regions by using the promoter from the gene encoding Ca2+/ calmodulin protein kinase IIalpha. In these R(AB) transgenic mice, hippocampal PKA activity was reduced, and L-LTP was significantly decreased in area CA1, without affecting basal synaptic transmission or the early phase of LTP. Moreover, the L-LTP deficit was paralleled by behavioral deficits in spatial memory and in long-term but not short-term memory for contextual fear conditioning. These deficits in long-term memory were similar to those produced by protein synthesis inhibition. Thus, PKA plays a critical role in the consolidation of long-term memory.

CREB is critical for long-lasting synaptic and behavioral plasticity in invertebrates. Its role in the mammalian hippocampus is less clear. We have interfered with CREB family transcription factors in region CA1 of the dorsal hippocampus. This impairs learning in the Morris water maze, which specifically requires the dorsal hippocampus, but not context conditioning, which does not. The deficit is specific to long-term memory, as shown in an object recognition task. Several forms of late-phase LTP are normal, but forskolin-induced and dopamine-regulated potentiation are disrupted. These experiments represent the first targeting of the dorsal hippocampus in genetically modified mice and confirm a role for CREB in hippocampus-dependent learning. Nevertheless, they suggest that some experimental forms of plasticity bypass the requirement for CREB.

In an attempt to improve behavioral memory, we devised a strategy to amplify the signal-to-noise ratio of the cAMP pathway, which plays a central role in hippocampal synaptic plasticity and behavioral memory. Multiple highfrequency trains of electrical stimulation induce long-lasting long-term potentiation, a form of synaptic strengthening in hippocampus that is greater in both magnitude and persistence than the short-lasting long-term potentiation generated by a single tetanic train. Studies using pharmacological inhibitors and genetic manipulations have shown that this difference in response depends on the activity of cAMPdependent protein kinase A. Genetic studies have also indicated that protein kinase A and one of its target transcription factors, cAMP response element binding protein, are important in memory in vivo. These findings suggested that amplification of signals through the cAMP pathway might lower the threshold for generating long-lasting long-term potentiation and increase behavioral memory. We therefore examined the biochemical, physiological, and behavioral effects in mice of partial inhibition of a hippocampal cAMP phosphodiesterase. Concentrations of a type IV-specific phosphodiesterase inhibitor, rolipram, which had no significant effect on basal cAMP concentration, increased the cAMP response of hippocampal slices to stimulation with forskolin and induced persistent long-term potentiation in CA1 after a single tetanic train. In both young and aged mice, rolipram treatment before training increased long-but not short-term retention in freezing to context, a hippocampus-dependent memory task.The second messenger, cAMP, and cAMP-dependent protein kinase A (PKA) have been implicated in short-and longlasting synaptic plasticity in Aplysia and in short-and longlasting behavioral learning in Aplysia and Drosophila (1, 2). Recently, convergent pharmacological and genetic evidence has also implicated the cAMP system in short-lasting longterm potentiation (LTP) at the mossy fiber-CA3 synapse of rodent hippocampus (3-6), and, strikingly, in the stronger longer-lasting intermediate and late phases of long-lasting LTP (L-LTP) that follow three to four trains of tetanic stimulation in all three hippocampal pathways: the perforant, the mossy fiber, and the Schaeffer collateral (CA3-CA1) (3,(7)(8)(9)(10)(11)(12)(13)(14). LTP is a well studied example of synaptic plasticity in mammals, thought to be a candidate cellular mechanism for mediating some forms of explicit hippocampus-dependent memory (15, 16). L-LTP has been of particular interest in regard to this behavioral correlation, because it is much more persistent than the short-lasting long-term potentiation that follows a single tetanic train (7,8,9). L-LTP persists as long as it has been observed, up to 29 hr in vitro, and depends at later time points not only on PKA activity but also on transcription and translation (3,6,8,9), much like behavioral long-term memory.The dependence of L-LTP, in hippocampal slices and behavioral memory, on PKA a...