Jump to ContentJump to Main Navigation
Brain, Perception, MemoryAdvances in Cognitive Neuroscience$

Johan J. Bolhuis

Print publication date: 2000

Print ISBN-13: 9780198524823

Published to Oxford Scholarship Online: March 2012

DOI: 10.1093/acprof:oso/9780198524823.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2020. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use.  Subscriber: null; date: 25 February 2020

Genetic strategies for the study of hippocampal-based memory storage

Genetic strategies for the study of hippocampal-based memory storage

(p.163) 10 Genetic strategies for the study of hippocampal-based memory storage
Brain, Perception, Memory

Danny G. Winder

Eric R. Kandel

Oxford University Press

Abstract and Keywords

This chapter is a review of recent genetic approaches to the study of learning and memory. With modern genetic techniques, individual genes can be ‘knocked out’ or, alternatively, overexpressed, and the effects of these manipulations on behaviour or on properties of neurons can be measured. This chapter discusses genetic approaches to long-term potentiation (LTP), where the genes coding for individual proteins that are thought to play a role in some stage of LTP can be deleted or overexpressed. Furthermore, it is important to realize that there is not a one-to-one relationship between a gene and behaviour. This chapter discusses ‘second generation’ genetic techniques, where there is a much higher degree of spatial and temporal specificity of the genetic manipulation. It is hoped that these more sophisticated genetic techniques, together with contemporary pharmacological approaches, can shed more light on the mechanisms underlying LTP, and also of those underlying memory.

Keywords:   genetic approaches, LTP, temporal specificity, genetic manipulation, pharmacological approaches


Memory is a remarkable mental faculty, one that is distinguished for both its importance and its variability. Under some circumstances, memory can persist reliably for decades. Under other circumstances, it can fade or become unreliable, or even distorted. The same person that might vividly recall in minute details the events surrounding their wedding 20 years ago can forget the details of a breakfast they had a few hours ago. Such variability in memory storage is important for our day-to-day lives. It is in part the variability of memory that helps in prioritizing the information that we do store. Indeed, a reasonable balance of memory storage and memory loss may be one of the essential requirements of a well-functioning free and independent mind. Memory storage that is either too detailed or too feeble would be detrimental to optimal functioning.

What then are the basic mechanisms of memory storage? Also what factors impart variability on the storage of memory? Since Ramon Y Cajal in his Croonian lecture to the Royal Society in 1894 first proposed that memory formation must involve a lasting modification in neuronal connections, neuroscientists (Gabriel Horn among them; see, for example, Horn, 1998, Chapter 19) have been pursuing the cellular substrate for learning and memory.

In the 1970s, Bliss and Lømo (1973) discovered long-term potentiation (LTP) of glutamatergic synaptic transmission. LTP is a robust enhancement of synaptic transmission, and currently is a leading candidate for the cellular substrate for hippocampal-based learning and memory in mammals. Thus, for the last 30 years, neuroscientists have been trying to unravel the molecular mechanisms underlying LTP and attempting to determine the relevance of LTP to memory storage (see Dudai and Morris, Chapter 9). In this chapter, we will briefly describe those aspects of LTP that make it an appealing candidate for a cellular substrate for memory formation. We then describe new genetic approaches for studying LTP that have begun to delineate the signal cascades necessary for distinct phases of LTP and the relationship of these cellular physiological phases to the phases of memory storage.

Basic Properties of LTP: A Cellular Substrate for Learning and Memory?

LTP is a lasting enhancement of synaptic transmission in response to a brief high-frequency stimulus that shares several properties with hippocampal-based learning and memory (Figure 10.1)

  1. (i) LTP occurs at synapses that, based on lesion studies, are thought to play a role in learning and memory. Studies ranging from rodents, monkeys and humans with temporal lobe damage all suggest that the hippocampus plays a key role in long-term memory formation. The hippocampus is a rudimentary cortical structure with a basic trisynaptic organization in which each of the main cell types is glutamatergic (Figure 10.1A). Each of the synapses in this structure can undergo LTP. The Schaffer collateral—CA1 pyramidal cell pathway is thought to be particularly important for memory storage, since selective lesions of this pathway in humans and in experimental animals produce significant interference in memory storage (Zola-Morgan et al., 1986)

  2. (ii) LTP is associative, like some forms of memory. This associativity is imparted by the unique properties of the N-methyl-D-aspartate (NMDA) (p.164)

                          Genetic strategies for the study of hippocampal-based memory storage

    Figure 10.1. (A) Hippocampal slice schematic including two stimulating electrodes and a central recording micropipet (adapted from a schematic kindly provided by Dr Robert Gereau, IV). (B) Typical fEPSPs recorded before and after the induction of LTP at the Schaffer collateral-CA1 pyramidal cell synapse in a mouse hippocampal slice. (C) Two-pathway experiment demonstrating the pathway specificity of LTP. Upward arrows in (C) and (D) represent timing of 1 s 100 Hz trains. (D) Multiple trains of tetanization recruit a non-decremental late phase of LTP.

    (p.165) subclass of glutamate receptor. Calcium influx through this receptor is required for NMDA receptor-dependent forms of LTP, yet this receptor is tonically blocked at resting membrane potentials by magnesium. Thus, the simultaneous presence of glutamate at postsynaptic NMDA receptors and postsynaptic depolarization is required to activate the NMDA receptor and to induce LTP.

  3. (iii) LTP is pathway specific (Figure 10.1C). One can record from one CA1 pyramidal cell and stimulate two independent inputs onto that cell that synapse on distinct populations of dendritic spines. If one of these pathways is tetanized so that it undergoes NMDA receptor-dependent LTP, the second pathway will not show LTP unless the synapses happen to be extremely close to one another (〈70 μm) (Engert and Bonhoeffer, 1997)

  4. (iv) LTP, like memory, exists in distinct temporal and biochemical phases (Huang et al., 1996). Thus, while short-term memory and short-lasting forms of LTP do not require protein or RNA synthesis, both long-term memory and a long-lasting form of LTP referred to as late-phase LTP require both RNA and protein synthesis (Figure 10.1D). These phases will be discussed in more detail later in this chapter.

  5. (v) Like memory, LTP is quite susceptible to modulation, so that stimuli which elicit little if any LTP in one context might elicit robust LTP in another context. For example, as we will discuss in more detail later, the neuromodulator noradrenaline can convert stimuli that normally never elicit LTP to stimuli that produce robust LTP. Paralleling these data, noradrenaline has been suggested to play facilitatory roles in learning and memory (McGaugh et al., Chapter 13).

Initial Approaches to the Study of LTP and the Link to Memory Storage

While these five properties of LTP make it an attractive candidate for a cellular substrate for memory storage, it has proven extremely difficult to test this idea. There are two key problems. First, LTP is not unique to the Schaffer collateral pathway, but can be elicited at many synaptic sites within the brain, making it difficult to know where to begin to study the relationship between LTP and memory. Even within the hippocampus, a structure strongly implicated in memory formation, LTP can be generated at each of the glutamatergic synapses in the hippocampal trisynaptic circuit. Secondly, there remains a great deal of controversy about mechanisms required for the initiation and maintenance of LTP.

The bulk of experiments assessing mechanisms underlying LTP, and the link between LTP and memory, have been pharmacological in nature until recently (e.g. Morris, 1989). Pharmacological approaches, while providing good temporal resolution, are limited in scope and specificity First, many molecules, such as transcription factors, that are interesting candidates as part of the LTP cascade are not tractable to pharmacological interventions. Secondly, pharmacological reagents often suffer from non-specific actions, which can be quite insidious in the study of LTP and memory. For example, one controversial aspect of the induction of LTP in the hippocampus is the possible involvement of metabotropic glutamate receptors (mGluRs). Several groups have demonstrated that mGluR antagonists reduce NMDA receptor-dependent LTP (Bashir et al., 1993; Riedel and Reymann, 1993; Richter-Levin et al., 1994). However, an approximately equal number of groups working under quite similar conditions have failed to replicate these findings (Bordi and Ugolini, 1995; Selig et al., 1995; Thomas and OʼDell, 1995). One possible explanation for this controversy was provided recently by Contractor et al. (1998) who demonstrated that the antagonists utilized in these studies bind not only to mGluRs but also to the glycine-binding site of the NMDA receptor (Contractor et al., 1998).

In this chapter, we will review a parallel genetic approach to the study of synaptic plasticity and memory that has emerged recently, and rapidly is becoming an extremely useful adjunct approach to pharmacological studies as well as allowing the study of targets that previously were inaccessible.

Genetic Approaches to the Study of LTP and Memory Storage

Studying LTP and memory storage with a genetic approach, deleting or overexpressing specific genes within an animal's genome, has a number of advantages, (i) This approach allows the study of some genes and their encoded proteins for which no pharmacological inhibitors exist. Moreover, in the (p.166) case where pharmacological inhibitors do exist, genetic approaches provide converging lines of data, (ii) By altering the expression of a single gene, genetic approaches provide precise and specific experimental manipulations, (iii) Mice in which genes have been knocked out or overexpressed can be bred to produce large numbers of identical genetically modified animals, greatly reducing experimental variability. (iv) Once created, genetically modified animals can be backcrossed onto different strains with distinctive characteristics (such as fast and slow learners). (v) Genetically modified mice allow direct correlations between biochemical, morphological and electrophysiological data on the one hand, and intact animal behavioral data on the other.

First Genetic Approaches to the study of LTP and Memory

Genetic approaches have had a major impact on the study of LTP and its relationship to hippocampal-based memory storage. For example, while it was clear from pharmacological studies that protein kinases are key signaling elements in LTP, it was not clear which particular kinases are involved. Leading candidates from early pharmacological studies included protein kinase C (PKC), calcium/calmodulin-dependent protein kinase IIα (CaMKIIα) and tyrosine kinases. To test these possibilities further, through targeted homologous recombination in embryonic stem (ES) cells, genes encoding specific isoforms of these kinases were deleted in mice. Briefly, the strategy for generating a knockout mouse involves the creation of a targeting vector that will undergo site-specific recombination in ES cells. ES cells that undergo this site-specific recombination, and thus have incorporated a portion of the targeting vector into the genome at the expense of a portion of the gene to be deleted, are then cloned and injected into blastocysts to produce chimeric animals that have a mosaic pattern of expression of the targeting vector. Chimeric animals that incorporate the targeting construct into the genome of stem cells are then mated to appropriate inbred mouse strains to produce non-mosaic mutant animals in which every cell in their body has the targeting construct integrated into the genome (for a more detailed description, see Soriano, 1995).

Studies from mice lacking the genes encoding CaMKIIα, PKC γ and the tyrosine kinase Fyn have provided critical converging data to those from pharmacological studies. Consistent with previous pharmacological studies, in mice lacking CaMKIIα:, virtually no NMDA receptor-dependent LTP could be evoked, suggesting that CaMKIIα: is an intrinsic signaling molecule in the minimal pathway necessary to elicit LTP (Silva et al., 1992b). Mice lacking Fyn had a dramatic reduction of LTP elicited by high-frequency tetanization similar to that seen in tyrosine kinase inhibitor experiments, but did not have deficits in NMDA receptor-dependent LTP induced by saturating postsynaptic depolarization (Grant et al., 1992). Interestingly, in mice lacking PKC γ, LTP was dramatically impaired, but could be restored through prior administration of prolonged low-frequency stimulation (Abeliovich et al., 1993a). Thus, these data suggest the possibility that Fyn and PKC γ play important regulatory roles in the induction of LTP, or are more critically involved in later stages of LTP. Consistent with a role in the regulation of the induction of LTP, recent studies with Fyn knockout mice show that tyrosine phosphorylation of the NR2A subunit of the NMDA receptor is reduced compared with wild-type mice (Tezuka et al., 1999).

Because these animals were deficient in hippocampal LTP, the performance of these animals was assessed on tasks that are thought to exploit hippocampal-dependent memory. One such task is the Morris water maze (Figure 10.2B). In this task, mice are placed in a pool of opaque water in which a hidden platform is positioned. The mice are trained over several trial blocks to find the hidden platform under conditions in which the location of the hidden platform and distal visual cues are held constant. Wild-type mice with an intact hippocampus can learn the location of the platform, as shown by a decreased latency to finding the platform and a greater than chance amount of time spent swimming in the quadrant that contained the platform when it is removed (for example, see left panel of Figure 10.2B). However, CaMKIIα and Fyn knockout mice were unable to perform this task successfully (Grant et al., 1992; Silva et al., 1992a). In contrast, PKC γ knockouts performed normally and were virtually indistinguishable from controls in Morris maze performance (Abeliovich et al., 1993b). Thus while the CaMKIIα and Fyn knockout data seem to support a role for LTP in spatial memory formation, the PKC γ knockouts do not, although the argument can be made that in vivo LTP may be elicited in these mice during maze training. Further, while the PKC γ knockout mice did not show deficits in performance on the Morris water maze, (p.167)

                      Genetic strategies for the study of hippocampal-based memory storage

Figures 10.2. Morris and Barnes mazes for the study of learning and memory in rodents. (A) Schematic of Barnes maze apparatus with examples of distinct search strategies employed by mice. (B) Schematic of Morris maze apparatus illustrating the similarities between the two mazes.

they performed poorly in fear-conditioning paradigms that are thought to require the hippocampus.

Limitations of traditional knockout techniques

Using the conventional knockout approach to create genetically modified mice has provided important new data both for the molecular basis of LTP and for the correlation between LTP and behavior. However, there are significant problems inherent in these genetic approaches. These include: (i) a lack of spatial control of the mutation; (ii) lack of temporal control of the mutation; (iii) the genetic background on which the mutation is made; and (iv) gene compensation.

  1. (i) In a conventional knockout, every cell in the animal's body lacks the gene of interest. Because of this, it is difficult to relate alterations in cell function in a particular region of the brain to whole animal behavior. It is possible, in fact likely, that unanticipated effects of the genetic alteration in other body regions could at least partially account for the behavioral phenotype. For example, a somatic or central nervous system (CNS) effect on motor coordination would seriously impair interpretation of learning and memory assays that require movement through space.

  2. (ii) In addition to each cell lacking the gene, each cell carries this modification for the lifetime of the organism. Thus, it becomes difficult to distinguish whether a particular phenotype observed in an adult animal is the result of the acute effect of that genetic modification in the adult animal or whether it is a consequence of the prolonged genetic alteration throughout development. For example, in the Fyn knockout mouse, in addition to the defect in LTP in the hippocampus, hippocampal anatomy was also altered, making it difficult to know what the locus of the LTP deficit is (see below).

  3. (iii) The genetic background of a strain of mice on which the genetic manipulation is made can contribute to the observed phenotype. Several (p.168) studies have demonstrated that different inbred strains of mice perform quite distinctly on various behavioral paradigms (Crawley et al., 1997). For example, two different mice lacking mGluR1 have been produced (Aiba et al., 1994; Conquet et al., 1994). While the mouse produced by Aiba et al. exhibits defective LTP at the Schaffer collateral—CA1 pyramidal cell synapse, the one produced by Conquet et al. is normal. Similarly, in two distinct lines of mice lacking the neurotrophin brain-derived neurotrophic factor (BDNF), deficits in LTP are seen in both lines, but deficits in basal synaptic transmission are seen in only one line (Korte et al., 1995; Patterson et al., 1996). Although there are other possibilities, one likely one is that the difference in phenotype represents different interactions between the mutation and the genetic background. To compound this problem, in many cases, genetically modified mice are bred as hybrids between different strains, resulting in a variable background from animal to animal.

  4. (iv) There is also the problem of compensation. The ability to delete selectively specific isoforms of various enzymes is both a blessing and a curse. In most cases, pharmacological tools are not available to discriminate the contributions of distinct isoforms of an enzyme such as a kinase, so targeted deletion represents dramatically increased specificity. However, in many cases, the effects of deleting one isozyme are obscured by the compensatory up-regulation of complementary isozymes, as in the case of the RIß protein kinase A (PKA) subunit knockout (Amieux et al., 1997).

Thus, while in combination with other approaches the data from these mice provide powerful converging lines of evidence for the role of particular molecules both in LTP and memory formation, additional genetic approaches are required to deal with these problems. One approach has been to use antisense technology or viral vectors to alter gene expression. While these techniques show promise, significant problems are associated with them that have slowed their development as techniques for altering gene expression in a temporally controllable fashion selectively within the CNS (for reviews, see Nicot and Pfaff, 1997; Wim et al., 1998). Our laboratories have therefore been interested in modifying genetic approaches in a variety of ways to compensate for the deficiencies which emerged from the first-generation studies of genetically modified mice.

Rescue of genetic defects by restoration of recombinant protein in knockout mouse

In an attempt to differentiate between chronic and acute effects of the lack of a given gene in a knockout animal, one potentially successful strategy would be to resupply the recombinant protein that the gene encodes. For example, as mentioned earlier, mice lacking the gene encoding the neurotrophin BDNF exhibit reduced LTP (Korte et al., 1995; Patterson et al., 1996) and basal synaptic transmission (Patterson et al., 1996). To determine whether this reflected a need for the protein during development of the hippocampus, Patterson et al. supplied recombinant BDNF to hippocampal slices from knockout mice. Interestingly, they found that pre-incubation of slices for 6–12 h with BDNF was sufficient to restore basal transmission and LTP back to wild-type levels. In an independent study, similar results were obtained by resupplying the BDNF gene by viral-mediated transfection (Korte et al., 1996). Thus, these data suggest that in the hippocampus BDNF has a relatively acute role in synaptic transmission and modification of that transmission.

Unfortunately, while this was a compelling strategy for the study of neurotrophins in the hippocampus, resupplying recombinant protein is only useful for a limited subset of genes that encode extracellular signaling molecules such as BDNF. Thus this strategy would not be viable in cases where intracellular proteins are targeted.

Overexpression of transgenes in mouse CNS

As an alternative to targeted deletion of genes of interest, another possibility is the overexpression of genes within the CNS. Briefly, in gene overexpression experiments, linearized cDNA constructs encoding the gene of interest are microinjected into pronuclei of fertilized eggs to allow for random insertion into the genome. These eggs are then injected into pseudo-pregnant females. Offspring from these mothers are then analyzed by Southern blot or polymerase chain reaction (PCR) to determine in which mice the DNA construct underwent random insertion, and these ‘founder’ mice are backcrossed to inbred mouse strains. Founder mice and their offspring express the transgene according to several factors, mainly determined (p.169) by the specific promoter utilized to drive transgene expression, the number of copies of the transgene that integrated into the genome and the chromosomal integration site (for a more detailed description, see Hogan et al., 1986).

There are several advantages to the use of gene overexpression. Historically, to determine which regions of the brain participate in memory formation, chemical or electrolytic lesions have been used. These lesion experiments, however, suffer from several problems contributed by the induction of the lesion, the lack of reversibility of the lesion and verification of the extent of the lesion. In addition, such lesions often destroy fibers of passage in the region of interest that are not necessarily involved integrally in the function of that nucleus. By overexpression, one potentially can express a gene within a specific brain region. This represents a great advantage over traditional lesion experiments in that axons of passage are not affected, and the region of interest need not be destroyed but, rather, synaptic plasticity can be up- or down-regulated selectively in that region. Secondly, by using promoters that drive expression postnatally, the developmental consequences of the genetic manipulation are less than those achieved with targeted recombination.

There are, however, several cautions that need to be exercised in interpreting data from random integration-mediated overexpression. For example, while a knockout is an all-or-none modification of the expression of a gene, overexpressed transgenes can vary widely in their level of expression from line to line, depending on the site of integration, the promoter used to direct transgene expression and the number of copies inserted. Secondly, random integration experiments may be affected differentially from line to line by insertion site artifacts. On the other hand, differences in the levels of expression between different lines allow comparisons of an allelic series which aids in the comparison of multiple lines of mice carrying the same transgene and in the clear determination of the extent to which the insertion site contributes to the phenotype. In contrast, knockout animals carry the same linked genes regardless of line, making this determination more difficult, but no less critical, as exhibited by the mGluRl and BDNF knockout mice.

In an effort to restrict the expression of transgenes spatially, several different promoters have been utilized to limit expression of transgenes to CNS alone, or to some subset of neural structures. One of the most successfully utilized promoters is the CaMKIIα promoter. This promoter is advantageous for two reasons. First, it limits expression of transgenes primarily to forebrain and the limbic system. Secondly, the expression of CaMKIIα is postnatal, reducing the problems associated with transgene effects on development.

With overexpression, one can generate mice that overexpress a wild-type form of a given gene, or a mutated form that acts as a dominant-negative or constitutively activated form. Mayford et al. were the first to use this promoter to overexpress a mutated, constitutively active form of CaMKIIα (CaMKIIT286) to test further the roles of CaMKIIα in LTP (Mayford et al., 1995). If activation of CaMKIIα is sufficient to induce LTP, then synaptic transmission should be saturated in mice overexpressing CaMKIIα. Mayford et al. actually found that while there was an approximate doubling of calcium-independent kinase activity in mutant mice expressing the transgene, LTP elicited by high-frequency stimulation was normal. However, overexpression of CaMKIIT286 shifted the frequency curve for stimuli that induce LTP versus long-term depression. In particular, 10 Hz stimuli near the endogenous theta frequency (3–7 Hz) firing range evokes LTP in wild-type animals but not in mutant mice. Thus these studies suggest that the roles of CaMKIIα in LTP may be more complex than was appreciated previously.

Although these animals showed no detectable decrease in LTP elicited by high-frequency stimulation traditionally used in the study of plasticity, they were nonetheless severely impaired in performance on a hippocampus-dependent task known as the Barnes maze (Barnes, 1979; Figure 10.2A). This task is a useful variant of the Morris water maze. The Barnes maze consists of a large circular platform around the periphery of which are spaced multiple holes. Under one of these holes is an escape chamber. Mice are trained to learn the location of the chamber by placing them in the center of the platform in a brightly lit room and applying a loud tone. These aversive stimuli compel the mouse to seek shelter. As in the Morris maze, in the spatial version of this task the position of the escape chamber and distal visual cues are kept constant across training trials. An advantage of this task compared with the Morris maze is the ease with which one can determine whether hippocampal- or non-hippocampal-based search strategies are employed. Bach et al. (1995) found that mice overexpressing the CaMKIIT286 transgene are selectively impaired in their ability to utilize the hippocampus-dependent spatial search strategy. Thus these studies (p.170) nicely illustrate that it is insufficient to evaluate LTP only at high frequencies of stimulation. One needs also to consider the effects of a genetic manipulation on LTP elicited by lower frequency, more physiologically relevant stimuli when comparing LTP phenotypes in the hippocampal slice with memory phenotypes in the awake behaving animal. Further, these findings demonstrated that overexpression of transgenes in the CNS with the CaMKIIα promoter is a useful means of obtaining information about the contribution of a particular gene to LTP.

Another use for overexpression strategies is to resupply a gene deleted through targeted recombination. For example, while mice lacking Fyn exhibited defective LTP, they also had quite dramatic anatomical defects in their hippocampi that could contribute to the LTP deficit. To examine this possibility, Kojima et al. (1997) overexpressed a Fyn transgene under the control of the CaMKIIα promoter in mice lacking the Fyn gene. Since the transgene was not expressed until day 10 postnatally, the anatomical defects in the Fyn knockouts were not rescued by Fyn transgene overexpression. However, LTP was much greater in the knockouts that expressed the transgene than in those that did not, suggesting a clear dissociation between the anatomical and electrophysiological defects of Fyn deletion.

Finally, overexpressed transgenes can be very useful as reporters of biological activity. For example, mice overexpressing the reporter gene β-galactosidase under the direction of either cAMP-regulated element (CRE) sites that can be bound by the transcription factor CRE-binding protein (CREB) or the GAP-43 promoter have been useful in determining the physiological stimuli that activate these promoters (Impey et al., 1996; Namgung et al., 1997). In addition, one could overexpress a protein tagged with green fluorescent protein to study subcellular targeting of that protein in real time.

Temporal control of genetic manipulations in the CNS

The overexpression technique suffers from many of the same problems as the traditional knockout technique, and a few unique ones. Thus, a second generation of genetically modified animals has begun to be employed. One attractive model is to regulate temporally the expression of transgenes incorporated through random insertion. While the CaMKIIα promoter provides some level of spatial and temporal restriction, one cannot turn the expression of the transgene on and off at will. Temporally regulated expression addresses two of the problems listed above. First, by being able to manipulate the timing of transgene expression, one can begin to differentiate between effects of a transgene during development versus acute effects in adult animals. Secondly, the reversal of a phenotype by turning off transgene expression argues against insertion site artifacts contributing to that phenotype. While several different inducible systems are in various stages of development, such as systems based on the insect ecdysone response (No et al., 1996) as well as tamoxifen-regulated systems (Feil et at, 1996; Zhang et al., 1996; Brocard et al., 1997; Schwenk et al., 1998), by far the most successfully utilized system for the inducible overexpression of transgenes in the adult mouse brain to date is the tetracycline transactivator system developed by Bujaard and colleagues (Figure 10.3).

Tetracycline transactivator system

The tetracycline transactivator system utilizes a tetracycline-controlled transactivator protein (tTA), containing the repressor from the tetracycline resistance operon and the transactivating domain of the viral protein VP16 (Furth et al., 1994). This mutant transactivator binds to the tet operator (tetO) and initiates transcription of the downstream gene. This system has the advantage that administration of tetracycline analogs displaces tTA from tetO, and thus suppresses transcription of the transgene. Since many tetracycline analogs have excellent bioavailability, this system is attractive as an inducible expression system in mouse. Thus, Mayford et al. generated mice overexpressing this transactivator using the CaMKIIα promoter. By crossing these mice with mice overexpressing the CaMKIIT286 transgene under the direction of tetO, they were able to obtain the same biochemical, electrophysiological and behavioral phenotypes observed in the first line of CaMKIIT286 mice, and demonstrate that the effects of the activated CaMKIIT286 transgene at each of these levels was reversed in mutant mice by administration of 1 mg/ml of the tetracycline analog doxycycline in the animals' drinking water for 2–3 weeks (Mayford et al., 1996). Thus these data demonstrate that the phenotype observed is not a function of an insertion artifact, and suggest that the effect on LTP is not due to the chronic presence of the transgene during animal development. This system has also been utilized (p.171)

                      Genetic strategies for the study of hippocampal-based memory storage

Figure 10.3. Second-generation gene overexpression strategies in mice: Schematic for the tTA system.

(p.172) recently to overexpress an activated form of calcineurin (Mansuy et al., 1998a; Winder et al., 1998). The phenotype of these mice, which will be discussed below, was also reversed by the administration of doxycycline in the animals' drinking water.

Reverse tetracycline transactivator system

While the tTA system is a major step forward in the development of inducible gene expression systems in mouse, it nonetheless still has weaknesses. For example, it would be advantageous to be able to turn a transgene on, rather than off, with the tetracycline analog. One could consider administering the analogs during gestation and then removing them during adulthood, but this must be done very carefully since tetracycline analogs in high doses can interfere with normal development, and can be taken up into bone where they then slowly leach out over an extended period of time (although see Chen et al., 1998).

In an attempt to circumvent some of these problems, Bujaard and colleagues performed chemical mutagenesis of tTA. One of these mutants, referred to as reverse tTA, or rtTA, had the unique feature that rather than tetracycline analogs inhibiting binding of the transactivator to tetO, these analogs are required to facilitate this binding (Kistner et al., 1996). Thus the rtTA system can, in principle, be utilized as a forward inducible system (Figure 10.4).

One of the problems originally associated with the rtTA system was leakiness (see, for example, No et al., 1996). Several laboratories have reported that there is a high level of basal expression of transgenes in the absence of tetracycline analogs in mice utilizing this system, and particularly in transient transfection systems. However, we recently have utilized the rtTA system successfully for overexpression of two different transgenes in the CNS, the reporter gene ß-galactosidase and the phosphatase calcineurin (Mansuy et al., 1998b). In these mice, we found that administration of doxycycline at 6 mg/kg in the animal's food for 2 weeks was sufficient to initiate expression of the transgene in both lines, and that removal of doxycycline for 2 weeks was sufficient to return the level of expression back to uninduced levels. Thus, through utilization of the rtTA system, one can obtain relatively rapid induction and reversal of transgene expression (although turnover rate is probably transgene specific), allowing for more creative experimental designs (see below). It is likely that the leakiness of the rtTA system seen in other reports is a product of positional effects of the insertion sites and, in the case of transient transfections, an unusually high number of copies inserted into the genome. One way to reduce these problems would be to sequence the insertion site in one of the lines of rtTA that does not give high levels of basal transgene expression, and ‘knock in’ rtTA into that locus using targeted recombination.

The Convergence of overexpression and knockouts:Cre-lox system for spatially restricted gene knockout

In addition to the inducible, spatially localized overexpression strategies we have already considered, it would be extremely useful to have similarly restricted methods for gene ablation. This possibility has become viable with the recent successful application of the CreloxP system in mouse brain (Figure 10.5). In this two-mouse system, loxP sites are inserted into the gene through standard targeting techniques in one line of mice. In the second line of mice, Cre recombinase is overexpressed under the direction of a tissue-specific promoter such as that for CaMKIIα. In double-positive progeny of the matings of these two mice, the Cre recombinase clips out DNA sequences between the loxP sites, generating a knockout. Through the utilization of specific promoters to drive expression of Cre recombinase, one can generate knockouts with some degree of temporal and spatial regulation. In the amazing extreme case, the Kandel and Tonegawa laboratories utilized the CaMKIIα promoter to drive expression of Cre recombinase, and found that in some lines of mice they were able to limit the knockout to the CA1 region. The Tonegawa laboratory then used this strategy to generate CA1-specific knockouts of the NMDAR1 subunit of the NMDA receptor that did not occur until about 3 weeks postnatally (Tsien et al., 1996a, b). In these animals, NMDA receptor-dependent LTP in area CA1 was abolished, while NMDA receptor-dependent LTP at the perforant path-dentate granule cell synapse was unaltered. Although the LTP deficit apparently was restricted to area CA1, there was nonetheless a dramatic impairment of performance on the spatial version of the Morris water maze. Since basal synaptic transmission and briefer, NMDA receptor-independent forms of plasticity were unaltered in these animals, these experiments are some of the strongest to date suggesting a direct link between NMDA receptor-dependent LTP in area CA1 of the hippocampus and learning and memory. (p.173)

                      Genetic strategies for the study of hippocampal-based memory storage

Figure 10.4. Second-generation gene overexpression strategies in mice: Schematic for the rtTA system.

                      Genetic strategies for the study of hippocampal-based memory storage

Figure 10.5. Second-generation Cre—lox-mediated gene knockout strategy in mice.

(p.175) Future directions with second-generation genetic models

The development of more temporally and spatially restricted genetic modifications now allows the pursuit of many avenues of investigation that would have been difficult with the first generation of genetically modified mice. For example, one can imagine administering doxycycline to brain slice preparations after dissection from mice expressing transgenes under the rtTA system to carry out intra-animal gene/no gene comparisons of LTP. Further, one could examine the effects of different levels of transgene expression by utilizing doxycycline dose-response curves. Indeed, Chen et al. have shown recently that several orders of magnitude lower concentrations of doxycycline (25 μg/ml in drinking water) can still fully suppress transgene expression in the CNS under the tTA system (Chen et al., 1998). This finding should greatly improve the kinetics of the tTA system by allowing more rapid reversal of the effects of doxycycline. Further, by using such a low dose, the toxic effects of doxycycline on development apparently are removed, allowing for experiments in which animals can be raised on doxycycline to suppress transgene expression through development. This should allow the development of mice overexpressing CRE recombinase under the tTA system to achieve temporally inducible knockouts in the CNS.

By utilizing the reversibility of regulatable systems, in particular the rtTA system, one can now begin to dissect the distinct components of learning and memory disrupted by a given transgene. For example, Mansuy et al. found that by overexpressing a calcineurin transgene in mouse brain, they could impair both LTP induced by two 100 Hz trains and performance on the spatial version of the Morris water maze (Figure 10.4; Mansuy et al., 1998b). However, mutant animals utilizing the rtTA system that were not administered doxycycline, and thus did not express the transgene, were able to navigate this task successfully. When these mice were now administered doxycycline for 2 weeks, and then tested again for their memory of the location of the hidden platform, they spent a chance amount of time in each of the four quadrants, while their control counterparts show a strong preference for the quadrant that contained the hidden platform. This apparent deficit in the retrieval of spatial information was reversible, since these mice could be removed from doxycycline and tested a third time, at which time they then again showed a memory of the location of the hidden platform.Thus these studies demonstrate that one effectively can manipulate the expression of the transgene during discrete periods of a behavioral experiment.

Utilization of Genetic Approaches for the Study of Signaling Cascades and Their Relevance to LTP, Neuromodulation and Memory Storage

With the array of genetic tools outlined above, neuroscientists are now in a position to test hypotheses that previously were very difficult to address. One area in which genetic approaches to the study of LTP are proving particularly useful is in the unraveling of signaling cascades that are required. Since it would be beyond the scope of this chapter to give a comprehensive review of the genetic approaches used to target signaling cascades involved in LTP and memory, we will focus on the genetic and pharmacological evidence suggesting the existence of multiple phases of NMDA receptor-dependent LTP and the implications for the function of the PKA signaling pathway in plasticity and memory.

Ltp at the schaffer collateral-CA1 pyramidal cell synapse exists in biochemically and stimulus-dependent distinct phases

As mentioned earlier, one of the attractive aspects of LTP as a cellular substrate of memory storage is that it exists in distinct temporal and biochemical phases. Thus, protein and RNA synthesis inhibitors reduce a late component of LTP elicited by strong stimulation, but do not affect LTP elicited by weak stimulation. For example, administration of a single 100 Hz train to the Schaffer collateral—CA1 pyramidal cell synapse results in a decremental form of LTP that lasts 2–3 h (Figure 10.1D). In contrast, administration of three to four 1 s 100 Hz trains spaced by 5 min elicits a non-decremen-tal form of LTP that lasts as long as a healthy slice preparation can be maintained. While the LTP elicited by one train is unaffected by protein and RNA synthesis inhibitors, LTP elicited by 3–4 trains is reduced 2–4 h after the tetanus by these inhibitors. This difference led to the discrimination between early and late phases of LTP (for a review, see Huang et al., 1996).

The finding that inhibitors of RNA synthesis reduce LTP elicited by multiple high-frequency trains (p.176) suggests that this LTP involves one or more transcription factors. In the invertebrates Aplysia and Drosophila, the transcription factor CREB has been found to play a critical role in the generation of long-term facilitation of synaptic transmission and long-term memory (for a review, see Abel and Kandel, 1998). Thus it has been suggested that CREB plays a role in late-phase LTP in hippocampus. Unfortunately, to date, knockout mice have provided ambiguous results, in large part due to isozyme compensation by other CREB isoforms.However, using a reporter mouse in which multiple CREs drive the expression of β-galactosidase, Impey et al. (1996) have demonstrated that stimuli that elicit late-phase LTP, but not early-phase LTP, elicit the initiation of CRE-dependent transcription.

What kinases initiate this transcriptionally dependent late phase of LTP? In the invertebrates Aplysia and Drosophila, previous studies have demonstrated a role for the PKA signaling cascade in long-term facilitation of synaptic transmission and long-term

                      Genetic strategies for the study of hippocampal-based memory storage

Figure 10.6. Cartoon of LTP elicted by phases underlying one, two and four trains of 100 Hz stimulation.

                      Genetic strategies for the study of hippocampal-based memory storage

Figure 10.7. Cartoon of putative intracellular mechanisms underlying distinct phases of LTP.

(p.178) memory, both of which are dependent on macromolecular synthesis. Consistent with a role for PKA in the late phase of LTP, pharmacological activation of the PKA signaling cascade elicits a macromolecular synthesis-dependent enhancement of transmission that is mutually occlusive with tetanus-evoked late-phase LTP (Frey et al., 1993). Conversely, inhibitors of PKA block the induction of late-phase LTP by repeated high-frequency tetanization. Just as with protein and RNA synthesis inhibitors in area CA1, PKA inhibitors were not effective in inhibiting LTP evoked by a single high-frequency train, but only LTP evoked by multiple high-frequency trains.

These data suggest that one of the distinctions between early- and late-phase LTP is that stronger stimuli that elicit late-phase LTP evoke the activation of additional second messenger systems (Figures 10.6 and 10.7). For example, a simple model for evoking more persistent LTP with repeated trains would be the stronger recruitment of a given kinase. If this were the case, it would be predicted that the kinase inhibitor would be more effective at inhibiting LTP evoked by a single high-frequency train, and less effective at inhibiting LTP evoked by repeated high-frequency trains. The fact that the opposite result was found with PKA inhibitors suggests that there is a threshold for the recruitment of PKA or a PKA-associated protein for the induction of late-phase LTP. Consistent with this idea, administration of a single high-frequency tetanus to hippocampal slices does not evoke the detectable activation of PKA, accumulation of cAMP or initiation of CRE-dependent transcription, while administration of multiple high-frequency trains does in an NMDA-receptor dependent manner (Chetkovich et al., 1991; Chetkovich and Sweatt, 1993; Frey et al., 1993; Roberson and Sweatt, 1996)

Data from genetically modified mice provide converging evidence that PKA has a role in late-phase LTP at the Schaffer collateral—CAl pyramidal cell synapse. Constitutive overexpression of a dominant-negative regulatory subunit of PKA [R(AB)] results in a modest reduction of total hippocampal PKA activity and a reduction in LTP elicited by four 100 Hz trains but not by a single 100 Hz train. Again, if the establishment of late-phase LTP was simply the result of the ramping up of PKA activity as the number of trains is increased, it would be predicted that LTP elicited by a single 100 Hz train would be reduced to a greater extent than LTP elicited by four 100 Hz trains.

In addition to providing genetic verification of pharmacological data, the R(AB) mice also allow comparisons to be made between the physiological phenotype and memory formation. Consistent with a proposed role for late-phase LTP in long-term memory storage, mice overexpressing the R(AB) transgene have deficient long-term memory, as they seem to learn spatial memory tasks such as the Morris water maze with the same time course as wild-types but, 24 h after the last training trial, probe trial analysis of memory in these animals showed that the R(AB) mutants could no longer remember the location of the hidden platform, while the wild-types did.

Interestingly, mice in which individual RIβ or CIβ subunits of the regulatory and catalytic subunits of PKA have been removed by targeted deletion do not show decreases in total PKA activity in hippocampal homogenates (Brandon et al., 1995; Qi et al., 1996). However, late-phase LTP was nonetheless reduced in area CA1 in the CIβ mice, although not to the degree seen in the R(AB) knockout mice. These findings suggest that specific isoforms of PKA play unique roles in LTP. Further, this provides an example of how, in some cases, overexpression of dominant-negative constructs can be a more effective strategy than targeted deletion, depending on the specific question being addressed.

Evidence for an intermediate phase intervening between the early and late phases

One interesting result that is not explained by a simple division of LTP into early and late phases at the Schaffer collateral—CAl pyramidal cell synapse is the finding that LTP evoked by multiple (3–4) high-frequency trains decays much more rapidly in the presence of PKA inhibitors than in the presence of protein synthesis inhibitors. This suggests that in addition to initiating a macromolecular synthesis-dependent late phase, PKA may participate in an intermediate, macromolecular synthesis-independent phase of LTP. A convergence of pharmacological and genetic data is consistent with this hypothesis. First, in contrast to LTP elicited by four 100 Hz trains, LTP elicited by two 100 Hz trains is independent of protein synthesis, yet still requires the activation of PKA. PKA inhibitors reduce LTP elicited by two 100 Hz trains in area CA1, and LTP induced by this protocol is reduced in R(AB) mice. Thus there exists a stimulus that can induce a form of PKA-dependent LTP that is macromolecular synthesis independent.

(p.179) What role does PKA play in this intermediate phase of LTP? A growing body of evidence suggests that phosphatases may play roles in constraining synaptic enhancement, and mediating long-term depression of transmission (Mulkey et al., 1993, 1994; Blitzer et al., 1995, 1998; Thomas et al., 1996; Winder et al., 1998). Several mechanisms have been proposed as determinants of the balance of activity between kinases and phosphatases. In particular, the Ca2+-activated phosphatase calcineurin (PP2B) has been a focus of attention because it is activated by calcium at concentrations that have little or no effect on Ca2+-activated kinases. In addition to potentially dephosphorylating key molecules mediating synaptic plasticity, calcineurin can dephosphorylate a family of small molecular weight proteins such as inhibitor-1 and DARPP-32 that regulate the activity of protein phosphatase-1 (PP1). PKA can inactivate this phosphatase cascade by phosphorylation of these small molecular weight proteins, resulting in the inhibition of PP1. Thus, it has been suggested that one role for PKA in LTP is to suppress phosphatase cascades to allow the full expression of LTP.

Both pharmacological and genetic evidence suggest that the role of PKA in this intermediate phase is to suppress an endogenous phosphatase cascade. First, inhibition of phosphatases pharmacologically removes the ability of PKA inhibitors to reduce LTP evoked by strong stimuli (Blitzer et al., 1995; Thomas et al., 1996). Secondly, injection into pyramidal neurons of mutated forms of the small molecular weight protein inhibitor-1, a point of convergence of PKA and the phosphatase calcineurin, disrupts LTP (Blitzer et al., 1998). Finally, overexpression of the phosphatase calcineurin, either constitutively with the CaMKIIα promoter or regulated with the tTA or rtTA systems, selectively inhibits PKA-dependent forms of LTP at the Schaffer collateral—CAl pyramidal cell synapse (Mansuy et al., 1998a, b; Winder et al., 1998). Interestingly, however, pharmacologically evoked late-phase LTP that bypasses the earlier phases was not decreased by overexpression of calcineurin. Thus, in total, a convergence of pharmacological data suggests that PKA plays multiple roles in mediating LTP through both macromolecular synthesis-independent and -dependent means (Figures 10.6 and 10.7). One of these roles appears to involve suppression by PKA of an endogenous phosphatase cascade to allow for the more effective phosphorylation of effector molecules by kinases such as CaMKIIα. A second role is to initiate a macromolecular synthesis-dependent long-lasting late-phase potentiation.

Multiple sources of cAMP: neuromodulation in the hippocampus

The apparent involvement of distinct kinases in distinct phases of LTP has potentially important implications. In particular, while CaMKIIα is activated primarily by influx of calcium from the extracellular space, the cAMP signaling cascade can be regulated by a myriad of stimuli, including Ca2+-, PKC, G-protein-coupled receptors, etc. Thus, it is conceivable that a wider variety of initiating stimuli could participate in the induction of PKA-dependent forms of LTP at the Schaffer collateral—CAl pyramidal cell synapse than in the NMDA receptor-dependent initiation of the early phase through activation of CaMKIIα. This is particularly interesting given previous studies showing that there are a large number of neuromodulatory inputs into the hippocampus and that the modulators can regulate learning and memory on the one hand, and cAMP levels on the other. In addition to intrinsic modulation of these glutamatergic synapses by NMDA and mGlu receptors, the hippocampus receives extrinsic neuromodulatory input from several sources. For example, it receives noradrenergic input from the locus ceruleus, serotonergic input from the dorsal raphe, cholinergic input from the lateral septal nucleus and dopaminergic input from the ventral tegmentum. Activation of receptors for these neuromodulatory inputs in the hippocampus regulates glutamatergic transmission through the hippocampal circuit through activation of G-proteins that regulate various second messenger cascades, including the cAMP signaling cascade.

Previous pharmacological studies have suggested that many neuromodulators important in memory storage also have a role in LTP in the hippocampus. Antagonists of D1/D5 dopamine receptors reduce the tetanus-induced increase in cAMP levels as well as LTP induced by these strong tetani (Frey et al., 1993; Huang and Kandel, 1995; Otmakhova and Lisman, 1996, 1998). In addition, a long-lasting increase in synaptic transmission that appears to share properties of late-phase LTP can be generated in a manner that bypasses electrical stimulation by application of agents that directly increase cAMP levels in area CA1, in particular D1/D5 dopamine receptor agonists (Huang and Kandel, 1995). Thus, these data suggest (p.180) that Dl /D5 dopamine receptors may play a role in the late phase of LTP by increasing cAMP levels.

In addition to dopaminergic receptor modulation of LTP, recent evidence suggests that β-adrenergic receptors can regulate LTP in a cAMP-dependent manner as well. For example, activation of the cAMP cascade by β-adrenergic receptor stimulation in area CA1 enables prolonged low-frequency stimulation, which normally elicits little change in synaptic transmission, to elicit robust LTP (Thomas et al., 1996; Katsuki et al., 1997, Winder et al., 1999).

CAMP as a key messenger for modulation of neuronal signaling

What is the particular role of PKA in these later phases of LTP? Although regulation of endogenous phosphatases is one apparent mechanism, as discussed below, the cAMP signaling cascade plays an important role in a number of aspects of hippocampal neuronal function, and its levels are regulated by a diverse array of mechanisms. In fact, as will be detailed below, increasing evidence exists for PKA-independent roles for cAMP in the hippocampus, necessitating the need for genetic approaches specifically to target PKA-dependent versus independent roles in LTP.

Rapid modulation of cellular excitability

Activation of the neuronal cAMP signaling cascade has almost exclusively excitatory effects on CA1 and CAS pyramidal cell excitability. Application of agents that enhance the cAMP signaling pathway at a gross level increases input resistance and to some degree depolarizes these neurons from their resting membrane potential. In addition, either directly or indirectly, cAMP elevations result in a decrease in the slow afterhyperpolarization (sAHP) that is thought to be responsible for limiting firing frequency, induce an inhibitory shift in the activation—inactivation kinetics of the A-current, enhance an inwardly rectifying chloride current, decrease the conductance of sodium channels, enhance L-type calcium channel availability and enhance currents elicited from P-type calcium channels (Madison and Nicoll, 1986; Gray and Johnston, 1987; Staley, 1994; Cantrell et al., 1997, Kavalali et al., 1997, Hoffman and Johnston, 1998). Data also suggest that cAMP may modulate hippocampal neuron excitability directly, without requiring PKA, through its actions on Ih and CNG channels (Pedarzani and Storm, 1995).

Rapid regulation of synaptic transmission

In addition to acute regulation of postsynaptic excitability, the cAMP signaling cascade can also regulate transmitter release from both glutamatergic and GABAergic synapses In the hippocampus. For example, activation of β-adrenergic receptors elicits a transient enhancement of synaptic transmission that is associated with an increased frequency of mini amplitudes but not frequency, and is associated with a decrease in paired-pulse facilitation (Gereau and Conn, 1994). Application of the adenylyl cyclase activator forskolin also results in a rapid enhancement of glutamatergic transmission that is; associated with an increase in mini frequency but not amplitude (Chavez-Noriega and Stevens, 1992). Similar results have been obtained with γ-aminobutyric acid (GABA) release (Capogna et al., 1995; Trudeau et al., 1996).

Current evidence suggests that multiple mechanisms may underlie the acute enhancement of synaptic transmission elicited by the cAMP signaling cascade. For example, both presynaptic calcium channels and synaptic vesicle release machinery are phosphorylated in response to activation of PKA (Parfitt et al., 1992; Hell et al., 1995).

Rapid regulation of postsynaptic responsiveness

Activation of the PKA signaling cascade can regulate postsynaptic responsiveness to released glutamate and GABA. Activation of this cascade selectively enhances glutamate responsiveness in cultured hippocampal neurons, probably through phosphorylation of GluR6, although similar results have not been obtained in slice preparations (Greengard et al., 1991; Wang et al., 1991). PKA can also regulate NMDA receptor function by reducing the desensitization elicited by calcineurin. Further, PKA phosphorylates a serine residue on the C-terminus of GluRl, and the phosphorylation state of this site is decreased after low-frequency stimulation that induces long-term depression of transmission (Kameyama et al., 1998). Finally, GABA responsiveness is reduced in a PKA-dependent manner in CA1 pyramidal cells (Poisbeau et al., 1999)

Future directions: genetic approaches to neuromodulation and cAMP-dependent plasticity

Given the large array of effects of the cAMP system on neuronal function, how does one determine which of these particular targets is important? The availability (p.181) of new genetic strategies should help in this regard. First, by expressing transgenes that regulate the cAMP cascade pre- and postsynaptically, one may be able to eliminate the contribution of one side of the synapse. Secondly, by genetically altering target molecules, one can tease out their contributions to PKA-dependent forms of LTP systematically. For example, if PKA-mediated phosphorylation of GluRl is important, then a ‘knock in’ of GluRl lacking the PKA phosphorylation site should impair PKA-dependent forms of LTP.

In addition, we need to know more about the temporal and spatial dynamics of the activation of the cAMP signaling cascade. While evidence suggests that tetanus-evoked increases in cAMP levels and PKA activity are transient in nature, it is likely that the increases in cAMP generated by the neuromodulators released during behavioral tasks could be more prolonged. Thus, one possibility is that tonic levels of PKA activity can influence PKA-dependent forms of LTP differently from transiently increased levels of cAMP. Recent experiments have shown that concentrations of a cAMP phosphodiesterase inhibitor that do not affect basal synaptic transmission of basal cAMP levels nonetheless facilitate cAMP responses evoked by forskolin, and LTP evoked by early-phase LTP-inducing stimuli (Barad et al., 1998). One could test this possibility genetically by comparing the effects of overexpression of an activated adenylyl cyclase or PKA with that of perhaps a dominant-negative form of a phosphodiesterase. Finally, given such a diverse array of possible effects elicited by the cAMP signaling cascade, the possibility must be considered that compartmentalized subcellular increases in cAMP levels may regulate distinct subsets of these effects. One approach to test this idea is to target these neuromodulatory systems genetically, for example by the generation of CAl-specific knockouts of the β 1 -adrenergic receptor and the D5 dopamine receptor, or by the overexpression of regulatable transgenes in noradrenaline- and dopamine-producing neurons that can be used to manipulate the endogenous release of these catecholamines. These types of genetically modified mice would then allow the determination of both the specific effects mediated by these modulators and the behavioral consequences of manipulation of these systems.


While the initial demonstration that mice in which kinases had been knocked out by targeted recombination enticed the scientific community with its remarkable specificity, the problems associated with those strategies at the time seemed severe and hard to overcome. In only a few years, however, significant progress has been made in addressing these problems. The speed with which this progress has been made suggests that further advances in these techniques are not far away Examination of the literature shows that we are already at a point where problems in LTP and memory storage can be seriously addressed with these techniques.


Bibliography references:

Abel, T. and Kandel, E.R. (1998) Positive and negative regulatory mechanisms that mediate long-term memory storage. Brain Research Reviews, 26, 360–378.

Abeliovich, A., Chen, C., Goda, Y., Silva, A.J., Stevens, C.F. and Tonegawa, S. (1993a) Modified hippocampal long-term potentiation in PKC γ-mutant mice. Cell, 75, 1253–1262.

Abeliovich, A., Paylor, R., Chen, C., Kim, J. J., Wehner, J.M. and Tonegawa, S. (1993b) PKC γ mutant mice exhibit mild deficits in spatial and contextual learning. Cell, 75, 1263–1271.

Aiba, A., Chen, C., Herrup, K., Rosenmund, C., Stevens, C.F. and Tonegawa, S. (1994) Reduced hippocampal long-term potentiation and context-specific deficit in associative learning in mGluRl mutant mice. Cell, 79, 365–375.

Amieux, P.S., Cummings, D.E., Motamed, K., Brandon, E.P., Wailes, L.A., Le, K., Idzerda, R.L. and McKnight, G.S. (1997) Compensatory regulation of Rla protein levels in protein kinase A mutant mice. Journal of Biological Chemistry, 272, 3993–3998.

Bach, M.E., Hawkins, R.D., Osman, M., Kandel, E.R. and Mayford, M. (1995) Impairment of spatial but not contextual memory in CaMKII mutant mice with a selective loss of hippocampal LTP in the range of the theta frequency. Cell, 81, 905–915.

Barad, M., Bourtchouladze, R., Winder, D.G., Golan, H. and Kandel, E.R. (1998) Rolipram, a type IV-specific phosphodiesterase inhibitor, facilitates the establishment of long-lasting long-term potentiation and improves memory. Proceedings of the National Academy of Sciences of the United States of America, 95, 15020–15025.

Barnes, C. (1979) Memory deficits associated with senescence: a neurophysiological and behavioral study in the rat. Journal of Comparative and Physiological Psychology, 93, 74–104.

Bashir, Z.I., Bortolotto, Z.A., Davies, C.H., Berretta, N., Irving, A.J., Seal, A.J., Henley, J.M., Jane, D.E., Watkins, J.C. and Collingridge, G.L. (1993) Induction of LTP in the hippocampus needs synaptic activation of glutamate metabotropic receptors. Nature, 363, 347–350.

Bliss, TV. and Lømo, T. (1973) Long-lasting potentiation of synaptic transmission in the dentate area of the anaesthetized rabbit following stimulation of the perforant path. Journal of Physiology, 232, 331–356.

(p.182) Blitzer, R.D., Wong, T., Nouranifar, R., Iyengar, R. and Landau, E.M. (1995) Postsynaptic cAMP pathway gates early LTP in hippocampal CA1 region. Neuron, 15, 1403–1414.

Blitzer, R.D., Connor, J.H., Brown, G.P., Wong, T., Shenolikar, S., Iyengar, R. and Landau, E.M. (1998) Gating of CaMKII by cAMP-regulated protein phosphatase activity during LTP. Science, 280, 1940–1942.

Bordi, F. and Ugolini, A. (1995) Antagonists of the metabotropic glutamate receptor do not prevent induction of long-term potentiation in the dentate gyrus of rats. European Journal of Pharmacology, 273, 291–294.

Brandon, E.R., Zhuo, M., Huang, Y.Y., Qi, M., Gerhold, K.A., Burton, K.A., Kandel, E.R., McKnight, G.S. and Idzerda, R.L. (1995) Hippocampal long-term depression and depotentiation are defective in mice carrying a targeted disruption of the gene encoding the RIβ subunit of cAMP-dependent protein kinase. Proceedings of the National Academy of Sciences of the United States of America, 92, 8851–8855.

Brocard, J., Warot, X., Wendling, O., Messaddeq, N., Vonesch, J.L., Chambon, P. and Metzger, D. (1997) Spatio-temporally controlled site-specific somatic mutagenesis in the mouse. Proceedings of the National Academy of Sciences of the United States of America, 94, 14559–14563.

Cantrell, A.R., Smith, R.D., Goldin, A.L., Scheuer, T., Catterall, W.A. (1997) Dopaminergic modulation of sodium current in hippocampal neurons via cAMP-dependent phosphorylation of specific sites in the sodium channel alpha subunit. Journal of Neuroscience, 17, 7330–7338.

Capogna, M., Gahwiler, B.H. and Thompson, S.M. (1995) Presynaptic enhancement of inhibitory synaptic transmission by protein kinases A and C in the rat hippocampus in vitro. Journal of Neuroscience, 15, 1249–1260.

Chavez-Noriega, L.E. and Stevens, C.F. (1992) Modulation of synaptic efficacy in field CA1 of the rat hippocampus by forskolin. Brain Research, 574, 85–92.

Chen, J., Kelz, M.B., Zeng, G., Sakai, N., Steffen, C., Shockett, P.E., Picciotto, M.R., Duman, R.S. and Nestler, E.J. (1998) Transgenic animals with inducible, targeted gene expression in brain. Molecular Pharmacology, 54, 495–503.

Chetkovich, D.M. and Sweatt, J.D. (1993) NMDA receptor activation increases cyclic AMP in area CA1 of the hippocampus via calcium/calmodulin stimulation of adenylyl cyclase. Journal of Neurochemistry, 61, 1933–1942.

Chetkovich, D.M., Gray, R., Johnston, D. and Sweatt, J.D. (1991) N-Methyl-D-aspartate receptor activation increases cAMP levels and voltage-gated Ca2+ channel activity in area CA1 of hippocampus. Proceedings of the National Academy of Sciences of the United States of America, 88, 6467–6471.

Conquet, E., Bashir, Z.I., Davies, C.H., Daniel, H., Ferraguti, E., Bordi, F. et al. (1994) Motor deficit and impairment of synaptic plasticity in mice lacking mGluRl. Nature, 372, 237–243.

Contractor, A., Gereau, R.W.T., Green, T. and Heinemann, S.F. (1998) Direct effects of metabotropic glutamate receptor compounds on native and recombinant N-methyl-D-aspartate receptors. Proceedings of the National Academy of Sciences of the United States of America, 95, 8969–8974.

Crawley, J.N., Belknap, J.K., Collins, A., Crabbe, J.C., Frankel, W., Henderson, N. et al. (1997) Behavioral phenotypes of inbred mouse strains: implications and recommendations for molecular studies. Psychopharmacology, 132, 107–124.

Engert, F. and Bonhoeffer, T. (1997) Synapse specificity of long-term potentiation breaks down at short distances. Nature, 388, 279–284.

Feil, R., Brocard, J., Mascrez, B., LeMeur, M., Metzger, D. and Chambon, P. (1996) Ligand-activated site-specific recombination in mice. Proceedings of the National Academy of Sciences of the United States of America, 93, 10887–10890.

Frey, U., Huang, Y.Y. and Kandel, E.R. (1993) Effects of cAMP simulate a late stage of LTP in hippocampal CA1 neurons. Science, 260, 1661–1664.

Furth, P.A., St Onge, L., Boger, H., Gruss, P., Gossen, M., Kistner, A., Bujard, H. and Hennighausen, L. (1994) Temporal control of gene expression in transgenic mice by a tetracycline-responsive promoter. Proceedings of the National Academy of Sciences of the United States of America, 91, 9302–9306.

Gereau, R.W.T. and Conn, P.J. (1994) Presynaptic enhancement of excitatory synaptic transmission by β-adrenergic receptor activation. Journal of Neurophysiology, 72, 1438–1442.

Grant, S.G., OʼDell, T.J., Karl, K.A., Stein, P.L., Soriano, P. and Kandel, E.R. (1992) Impaired long-term potentiation, spatial learning, and hippocampal development in fyn mutant mice. Science, 258, 1903–1910.

Gray, R. and Johnston, D. (1987) Noradrenaline and β-adrenoceptor agonists increase activity of voltage-dependent calcium channels in hippocampal neurons. Nature, 327, 620–622.

Greengard, P., Jen, J., Nairn, A.C. and Stevens, C.F. (1991) Enhancement of the glutamate response by cAMP-dependent protein kinase in hippocampal neurons. Science, 253, 1135–1138.

Hell, J.W., Yokoyama, C.T., Breeze, L.J., Chavkin, C. and Catterall, W.A. (1995) Phosphorylation of presynaptic and postsynaptic calcium channels by cAMP-dependent protein kinase in hippocampal neurons. EMBO Journal, 14, 3036–3044.

Hoffman, D.A., Johnston, D. (1998) Down-regulation of transient K+ channels in dendrites of hippocampal CA1 pyramidal neurons by activation of PKA and PKC. Journal of Neuroscience, 18, 3521–3528.

Hogan, B., Constantini, F. and Lacy, E. (1986) Manipulating the Mouse Embryo: A Laboratory Manual. Cold Spring Harbor Press, Cold Spring Harbor, New York.

Horn, G. (1998) Visual imprinting and the neural mechanisms of recognition memory. Trends in Neurosciences, 21, 300–305.

Huang, Y.Y. and Kandel, E.R. (1995) D1/D5 receptor agonists induce a protein synthesis-dependent late potentiation in the CA1 region of the hippocampus. Proceedings of the National Academy of Sciences of the United States of America, 92, 2446–2450.

Huang, Y.-Y., Nguyen, P.V., Abel, T. and Kandel, E.R. (1996) Long-lasting forms of synaptic potentiation in the mammalian hippocampus. Learning and Memory, 3, 74–85.

Impey, S., Mark, M., Villacres, E.C., Poser, S., Chavkin, C. and Storm, D.R. (1996) Induction of CRE-mediated gene expression by stimuli that generate long-lasting LTP in area CA1 of the hippocampus. Neuron, 16, 973–982.

(p.183) Kameyama, K., Lee, H.-K., Bear, M.F. and Huganir, R.L. (1998) Involvement of a postsynaptic protein kinase A substrate in the expression of homosynaptic long-term depression. Neuron, 21, 1163–1175.

Katsuki, H., Izumi, Y. and Zorumski, C.F (1997) Noradrenergic regulation of synaptic plasticity in the hippocampal CA1 region. Journal of Neurophysiology, 77, 3013–3020.

Kavalali, E.T., Hwang, K.S. and Plummer, M.R. (1997) cAMP-dependent enhancement of dihydropyridine-sensitive calcium channel availability in hippocampal neurons. Journal ofNeuroscience, 17, 5334–5348.

Kistner, A., Gossen, M., Zimmermann, F., Jerecic, J., Ullmer, C., Lubbert, H. and Bujard, H. (1996) Doxycycline-mediated quantitative and tissue-specific control of gene expression in transgenic mice. Proceedings of the National Academy of Sciences of the United States of America, 93, 10933–10938.

Kojima, N., Wang, J., Mansuy, I.M., Grant, S.G., Mayford, M. and Kandel, E.R. (1997) Rescuing impairment of long-term potentiation in fyn-deficient mice by introducing Fyn transgene. Proceedings of the National Academy of Sciences of the United States of America, 94, 4761–4765.

Korte, M., Carroll, P., Wolf, E., Brem, G., Thoenen, H. and Bonhoeffer, T. (1995) Hippocampal long-term potentiation is impaired in mice lacking brain-derived neurotrophic factor. Proceedings of the National Academy of Sciences of the United States of America, 92, 8856–8860.

Korte, M., Griesbeck, O., Gravel, C., Carroll, P., Staiger, V., Thoenen, H. and Bonhoeffer, T. (1996) Virus-mediated gene transfer into hippocampal CA1 region restores long-term potentiation in brain-derived neurotrophic factor mutant mice. Proceedings of the National Academy of Sciences of the United States of America, 93, 12547–12552.

Madison, D.V. and Nicoll, R.A. (1986) Cyclic adenosine 3′, 5′-monophosphate mediates ß-receptor actions of noradrenaline in rat hippocampal pyramidal cells. Journal of Physiology (London), 372, 245–259.

Mansuy, I.M., Mayford, M., Jacob, B., Kandel, E.R. and Bach, M.E. (1998a) Restricted and regulated overexpression reveals calcineurin as a key component in the transition from short-term to long-term memory. Cell, 92, 39–49.

Mansuy, I.M., Winder, D.G., Moallem, T.M., Osman, M., Mayford, M., Hawkins, R.D. and Kandel, E.R. (1998b) Inducible and reversible gene expression with the rtTA system for the study of memory. Neuron, 21, 257–265.

Mayford, M., Bach, M.E., Huang, Y.Y., Wang, L., Hawkins, R.D. and Kandel, E.R. (1996) Control of memory formation through regulated expression of a CaMKII transgene. Science, 274, 1678–1683.

Mayford, M., Wang, J., Kandel, E.R. and OʼDell, T.J. (1995) CaMKII regulates the frequency-response function of hippocampal synapses for the production of both LTD and LTP. Cell, 81, 891–904.

Morris, R.G. (1989) Synaptic plasticity and learning: selective impairment of learning in rats and blockade of long-term potentiation in vivo by the N-methyl-D-aspartate receptor antagonist AP5. Journal ofNeuroscience, 9, 3040–3057.

Mulkey, R.M., Herron, C.E. and Malenka, R.C. (1993) An essential role for protein phosphatases in hippocampal long-term depression. Science, 261, 1051–1055.

Mulkey, R.M., Endo, S., Shenolikar, S. and Malenka, R.C. (1994) Involvement of a calcineurin / inhibitor-1 phosphatase cascade in hippocampal long-term depression. Nature, 369, 486–488.

Namgung, U., Matsuyama, S. and Routtenberg, A. (1997) Long-term potentiation activates the GAP-43 promoter: selective participation of hippocampal mossy cells. Proceedings of the National Academy of Sciences of the United States of America, 94, 11675–11680.

Nicot, A. and Pfaff, D.W. (1997) Antisense oligodeoxynucleotides as specific tools for studying neuroendocrine and behavioral functions: some prospects and problems. Journal ofNeuroscience Methods, 71, 45–53.

No, D., Yao, TP. and Evans, R.M. (1996) Ecdysone-inducible gene expression in mammalian cells and transgenic mice. Proceedings of the National Academy of Sciences of the United States of America, 93, 3346–3351.

Otmakhova, N.A. and Lisman, J.E. (1996) D1/D5 dopamine receptor activation increases the magnitude of early long-term potentiation at CA1 hippocampal synapses. Journal of Neuroscience, 16, 7478–7486.

Otmakhova, N.A. and Lisman, J.E. (1998) D1/D5 dopamine receptors inhibit depotentiation at CA1 synapses via cAMP-dependent mechanism. Journal of Neuroscience, 18, 1270–1279.

Parfitt, K.D., Doze, V.A., Madison, D.V. and Browning, M.D. (1992) Isoproterenol increases the phosphorylation of the synapsins and increases synaptic transmission in dentate gyrus, but not in area CA1, of the hippocampus. Hippocampus, 2, 59–64.

Patterson, S.L., Abel, T., Deuel, T.A., Martin, K.C., Rose, J.C. and Kandel, E.R. (1996) Recombinant BDNF rescues deficits in basal synaptic transmission and hippocampal LTP in BDNF knockout mice. Neuron, 16, 1137–1145.

Pedarzani, P. and Storm, J.E (1995) Protein kinase A-independent modulation of ion channels in the brain by cyclic AMP. Proceedings of the National Academy of Sciences of the United States of America, 92, 11716–11720.

Poisbeau, P., Cheney, M.C., Browning, M.D. and Mody, I. (1999) Modulation of synaptic GABAA receptor function by PKA and PKC in adult hippocampal neurons. Journal of Neuroscience, 19, 674–683.

Qi, M., Zhuo, M., Skalhegg, B.S., Brandon, E.R, Kandel, E.R., McKnight, G.S. and Idzerda, R.L. (1996) Impaired hippocampal plasticity in mice lacking the Cßl catalytic subunit of cAMP-dependent protein kinase. Proceedings of the National Academy of Sciences of the United States of America, 93, 1571–1756.

Richter-Levin, G., Errington, M.L., Maegawa, H. and Bliss, TV. (1994) Activation of metabotropic glutamate receptors is necessary for long-term potentiation in the dentate gyrus and for spatial learning. Neuropharmacology, 33, 853–857.

Riedel, G. and Reymann, K. (1993) An antagonist of the metabotropic glutamate receptor prevents LTP in the dentate gyrus of freely moving rats. Neuropharmacology, 32, 929–931.

Roberson, E.D. and Sweatt, J.D. (1996) Transient activation of cyclic AMP-dependent protein kinase during hippocampal long-term potentiation. Journal of Biological Chemistry, 271, 30436–30441.

Schwenk, E., Kuhn, R., Angrand, P.O., Rajewsky, K. and Stewart, A.F. (1998) Temporally and spatially regulated (p.184) somatic mutagenesis in mice. Nucleic Acids Research, 26, 1427–1432.

Selig, D.K., Lee, H.K., Bear, M.F. and Malenka, R.C. (1995) Reexamination of the effects of MCPG on hippocampal LTP, LTD, and depotentiation. Journal of Neurophysiology, 74, 1075–1082.

Silva, A.J., Paylor, R., Wehner, J.M. and Tonegawa, S. (1992a) Impaired spatial learning in α-calcium-calmodulin kinase II mutant mice. Science, 257, 206–211.

Silva, A.J., Stevens, C.F., Tonegawa, S. and Wang, Y. (1992b) Deficient hippocampal long-term potentiation in a-calcium-calmodulin kinase II mutant mice. Science, 257, 201–206.

Soriano, P. (1995) Gene targeting in ES cells. Annual Review of Neuroscience, 18, 1–18.

Staley, K. (1994) The role of an inwardly rectifying chloride conductance in postsynaptic inhibition. Journal of Neurophysiology, 72, 273–284.

Tezuka, T., Umemori, H., Akiyama, T., Nakanishi, S. and Yamamoto, T. (1999) PSD-95 promotes Fyn-mediated tyrosine phosphorylation of the N-methyl-D-aspartate receptor subunit NR2A. Proceedings of the National Academy of Sciences of the United States of America, 96, 435–440.

Thomas, M.J. and OʼDell, T.J. (1995) The molecular switch hypothesis fails to explain the inconsistent effects of the metabotropic glutamate receptor antagonist MCPG on long-term potentiation. Brain Research, 695, 45–52.

Thomas, M.J., Moody, T.D., Makhinson, M. and OʼDell, T.J. (1996) Activity-dependent β-adrenergic modulation of low frequency stimulation induced LTP in the hippocampal CA1 region. Neuron, 17, 475–482.

Trudeau, L.E., Emery, D.G. and Haydon, P.G. (1996) Direct modulation of the secretory machinery underlies PKA-dependent synaptic facilitation in hippocampal neurons. Neuron, 17, 789–797.

Tsien, J.Z., Chen, D.F., Gerber, D., Tom, C., Mercer, E.H., Anderson, D.J., Mayford, M., Kandel, E.R. and Tonegawa, S. (1996a) Subregion- and cell type-restricted gene knockout in mouse brain. Cell, 87, 1317–1326.

Tsien, J.Z., Huerta, P.T. and Tonegawa, S. (1996b) The essential role of hippocampal CA1 NMDA receptor-dependent synaptic plasticity in spatial memory. Cell, 87, 1327–1338.

Wang, L.Y., Salter, M.W. and MacDonald, J.R (1991) Regulation of kainate receptors by cAMP-dependent protein kinase and phosphatases. Science, 253, 1132–5.

Wim, T.J., Hermens, M.C. and Werhaagen, J. (1998) Viral vectors, tools for gene transfer in the nervous system. Progress in Neurobiology, 55, 399–432.

Winder, D.G., Mansuy, I.M., Osman, M., Moallem, T.M. and Kandel, E.R. (1998) Genetic and pharmacological evidence for a novel, intermediate phase of long-term potentiation suppressed by calcineurin. Cell, 92, 25–37.

Winder, D.G., Martin, K.C., Muzzio, LA., Rohrer, D., Chruscinski, A., Kobilka, B. Kandel, E.R. (1999) ERK plays a regulatory role in induction of LTP by theta frequency stimulation and its modulation by beta-adrenergic receptors. Neuron, 24, 715–726.

Zhang, Y., Riesterer, C., Ayrall, A.M., Sablitzky, R., Littlewood, T.D. and Reth, M. (1996) Inducible site-directed recombination in mouse embryonic stem cells. Nucleic Acids Research, 24, 543–548.

Zola-Morgan, S., Squire, L.R. and Amaral, D.G. (1986) Human amnesia and the medial temporal region: enduring memory impairment following a bilateral lesion limited to field CA1 of the hippocampus. Journal of Neuroscience, 6, 2950–2967.