CREB and the formation of long-term memory

Cyclic AMP response element binding protein (CREB)-responsive transcription plays a central role in the formation of long-term memory in Drosophila, Aplysia and mice. Agents that disrupt the activity of CREB specifically block the formation of long-term memory, whereas agents that increase the amount or activity of the transcription factor accelerate the process. These results have led to the recent hypothesis that CREB is pivotal in the switch from short-term (protein synthesis independent) to long-term (protein synthesis dependent) memory.

Effects of a conditional Drosophila PKA mutant on olfactory learning and memory

The requirement for cAMP-dependent protein kinase (PKA) in associative learning of Drosophila was assessed in mutant flies hemizygous for a cold-sensitive allele, X4, of the DC0 gene, which encodes the major catalytic subunit of PKA. DC0X4 hemizygotes died as third-instar larvae at 18 degrees C, the restrictive temperature, but were viable when raised at 25 degrees C. Shifting adult DC0X4 hemizygotes from 25 degrees C to 18 degrees C led to a decrease in PKA activity from 24% to 16% of wild-type without impairing viability. At 25 degrees C, DC0X4 hemizygotes exhibited reduced initial learning relative to controls but normal memory decay in a Pavlovian olfactory learning assay. Shifting the temperature from 25 degrees C to 18 degrees C prior to training reduced initial learning to a similar extent in DC0X4 hemizygotes and controls but resulted in a steeper memory decay curve only in DC0X4 hemizygotes. These observations are suggestive of a role for PKA in medium-term memory formation in addition to its previously established role in initial learning.

Molecular cloning of linotte in Drosophila: a novel gene that functions in adults during associative learning

The linotte (lio) gene was identified in a screen for mutations that disrupted 3 hr memory after olfactory associative learning, without affecting the perception of odors or electroshock. The mutagenesis yielded a transposon-tagged gene disruption, which allowed rapid cloning of genomic DNA. The lio transcription unit was identified via rescue of the lio1 learning/memory defect by induced expression of a lio+ transgene in adults. The perception of odors or electroshock remained normal when the lio+ transgene was expressed in these lio1 flies. Learning/memory remained normal when the lio+ transgene was expressed in wild-type (lio+) flies. The lio gene produces only one transcript, the level of expression of which varies throughout development. Sequence analysis indicates that lio encodes a novel protein.

Molecular and biochemical characterization of dNOS: a Drosophila Ca2+/calmodulin-dependent nitric oxide synthase

Nitric oxide (NO) is an intercellular messenger involved with various aspects of mammalian physiology ranging from vasodilation and macrophage cytotoxicity to neuronal transmission. NO is synthesized from L-arginine by NO synthase (NOS). Here, we report the cloning of a Drosophila NOS gene, dNOS, located at cytological position 32B. The dNOS cDNA encodes a protein of 152 kDa, with 43% amino acid sequence identity to rat neuronal NOS. Like mammalian NOSs, DNOS protein contains putative binding sites for calmodulin, FMN, FAD, and NADPH. DNOS activity is Ca2+/calmodulin dependent when expressed in cell culture. An alternative RNA splicing pattern also exists for dNOS, which is identical to that for vertebrate neuronal NOS. These structural and functional observations demonstrate remarkable conservation of NOS between vertebrates and invertebrates.

A Drosophila CREB/CREM homolog encodes multiple isoforms, including a cyclic AMP-dependent protein kinase-responsive transcriptional activator and antagonist

We have characterized a Drosophila gene that is a highly conserved homolog of the mammalian cyclic AMP (cAMP)-responsive transcription factors CREB and CREM. Uniquely among Drosophila genes characterized to date, it codes for a cAMP-responsive transcriptional activator. An alternatively spliced product of the same gene is a specific antagonist of cAMP-inducible transcription. Analysis of the splicing pattern of the gene suggests that the gene may be the predecessor of the mammalian CREB and CREM genes.

Dissection of memory formation: from behavioral pharmacology to molecular genetics

Behavioral pharmacology has suggested an intricate, multiphasic pathway of memory consolidation. An integrated molecular pharmacological approach in Drosophila has lent support to this theory recently by dissecting consolidated memory into two genetically distinct components: a cycloheximide-insensitive, anesthesia-resistant memory and a cycloheximide-sensitive long-term memory. In addition, experiments using inducible dominant-negative transgenes in Drosophila or gene knockouts in mice demonstrate a role for cAMP-responsive transcription factors in formation of long-term memory. These studies support the application of reverse-genetic strategies, including the use of temporally specific agonists and antagonists, to advance the functional dissection of memory formation.

CREB as a memory modulator: induced expression of a dCREB2 activator isoform enhances long-term memory in Drosophila

Genetic studies of memory formation in Drosophila have revealed that the formation of a protein synthesis-dependent long-term memory (LTM) requires multiple training sessions. LTM is blocked specifically by induced expression of a repressor isoform of the cAMP-responsive element-binding protein (CREB). Here, we report an enhancement of LTM formation after induced expression of an activator isoform of dCREB2. Maximum LTM is achieved after one training session, and its formation depends on phosphorylation of the activator transgene. A model of LTM formation based on differential regulation of CREB isoforms is proposed.

Genetic dissection of consolidated memory in Drosophila

Behavioral and pharmacological experiments in many animal species have suggested that memory is consolidated from an initial, disruptable form into a long-lasting, stable form within a few hours after training. We combined these traditional approaches with genetic analyses in Drosophila to show that consolidated memory of conditioned (learned) odor avoidance 1 day after extended training consisted of two genetically distinct, functionally independent memory components: anesthesia-resistant memory (ARM) and long-term memory (LTM). ARM decayed away within 4 days, was resistant to hypothermic disruption, was insensitive to the protein synthesis inhibitor cycloheximide (CXM), and was disrupted by the radish single-gene mutation. LTM showed no appreciable decay over 7 days, was sensitive to CXM, and was not disrupted by the radish mutation.

Induction of a dominant negative CREB transgene specifically blocks long-term memory in Drosophila

Consolidated memory after olfactory learning in Drosophila consists of two components, a cycloheximide-sensitive, long-term memory (LTM) and a cycloheximide-insensitive, anesthesia-resistant memory (ARM). Using an inducible transgene that expresses a dominant negative member of the fly CREB family, LTM was specifically and completely blocked only after induction, while ARM and learning were unaffected. These results suggest that LTM formation requires de novo gene expression probably mediated by CREB family genes.

Memory through metamorphosis in normal and mutant Drosophila

To establish that a stable, long-lasting form of memory exists in Drosophila, we trained third-instar larvae by electroshocking them in the presence of a specific odor using a Pavlovian conditioning procedure. We show that conditioned odor avoidance produced in larvae still was present in adults 8 d later. Such memory through metamorphosis was specific to the temporal pairing of odor and shock; presentations of odors alone or shock alone did not produce a change. Thus, the memory involved associative processes. We also show that similar training of the single-gene memory mutants dunce and amnesiac did not yield any detectable learning in larvae or memory retention in adults, suggesting that these mutations interfere with long-term memory (LTM) formation even if LTM is induced independently of earlier memory retention processes.

Differential effects of dunce mutations on associative learning and memory in Drosophila

Initial learning, 30- and 180-min memory retention after Pavlovian conditioning of an odor avoidance response was quantified in dnc1, dnc2, dncM11 and Canton-S (wild-type) homozygotes and in dnc1/FM7, dnc2/FM7, dncM11/FM7, dncM11/Can-S, Can-S/FM7, dnc1/dncM11 and dnc2/dncM11 heterozygotes. Our results consistently showed that a) the dunce mutations are semi-dominant for initial learning and b) genetic variants carrying the enzymatically hypomorphic dnc2 mutation produce learning scores lower than those of the amorphic dncM11. Analysis of this particular set of retention intervals, using a modified statistical model designed to evaluate decay rates, revealed no discernable effects of the dunce mutations on memory formation 30 to 180 min after training. These results are consistent with a model of memory formation, in which dunce is hypothesized to disrupt acquisition and/or short-term memory.

Identification of linotte, a new gene affecting learning and memory in Drosophila melanogaster

We describe the identification of linotte, a new autosomal gene in Drosophila involved with learning and memory. The linotte mutant was derived from a PlacW transposon mutagenesis and was screened for three-hour memory deficits after classical conditioning of an olfactory avoidance response. Sensory and motor systems (olfactory acuity and shock reactivity) required for the classical conditioning experiments were normal in mutant linotte flies--indicating that the mutation disrupts learning/memory specifically. A chromosomal deficiency of the 37D region, where the linotte P insert was localized in situ, failed to complement linotte's memory defect, and flies from two lines homozygous for independent PlacW excisions show normal memory--indicating that the P insertion is responsible for the mutant phenotype. Additional behavior-genetic data suggest that linotte gene is non-vital.

Human amyloid precursor protein ameliorates behavioral deficit of flies deleted for Appl gene

Drosophila amyloid precursor protein-like (Appl) gene encodes a protein product (APPL) similar to beta-amyloid precursor protein (APP) associated with Alzheimer's disease. To understand the in vivo function of APPL protein, we have generated flies deleted for the Appl gene. These flies are viable, fertile, and morphologically normal, yet they exhibit subtle behavioral deficits. We show that a fast phototaxis defect in Appl- flies is partially rescued by transgenes expressing the wild-type, but not a mutant, APPL protein. We further demonstrate a functional homology between APPL and APP, since transgenes expressing human APP show a similar level of rescue as transgenes expressing fly APPL.

latheo, a new gene involved in associative learning and memory in Drosophila melanogaster, identified from P element mutagenesis

Genetic dissection of learning and memory in Drosophila has been limited by the existence of ethyl methanesulfonate (EMS)-induced mutations in only a small number of X-linked genes. To remedy this shortcoming, we have begun a P element mutagenesis to screen for autosomal mutations that disrupt associative learning and/or memory. The generation of "P-tagged" mutant alleles will expedite molecular cloning of these new genes. Here, we describe a behavior-genetic characterization of latheoP1, a recessive, hypomorphic mutation of an essential gene. latheoP1 flies perform poorly in olfactory avoidance conditioning experiments. This performance deficit could not be attributed to abnormal olfactory acuity or shock reactivity-two task-relevant "peripheral" behaviors which are used during classical conditioning. Thus, the latheoP1 mutation appears to affect learning/memory specifically. Consistent with chromosomal in situ localization of the P element insertion, deficiencies of the 49F region of the second chromosome failed to complement the behavioral effect of the latheoP1 mutation. Further complementation analyses between latheoP1 and lethal alleles, produced by excision of the latheoP1 insert or by EMS or gamma-rays, in the 49F region mapped the latheo mutation to one vital complementation group. Flies heterozygous for latheoP1 and one of two EMS lethal alleles or one lethal excision allele also show the behavioral deficits, thereby demonstrating that the behavioral and lethal phenotypes co-map to the same locus.

Reassessment of the effect of biological rhythm mutations on learning in Drosophila melanogaster

A link between learning deficits and circadian period-lengthening mutations in Drosophila melanogaster previously has been reported. Mutant long-period males performed poorly in two learning assays involving experience-dependent courtship inhibition. In one, normal males that have courted fertilized females subsequently show courtship inhibition with virgin females. In the other, normal males that have courted sexually immature males subsequently fail to court other immature males. Those results have been reassessed in an extended study of genetic variants involving the period gene. 1. Long-period perL1 males demonstrated poor conditioned courtship inhibition when exposed to fertilized females; they showed normal courtship conditioning when exposed to immature males. This could be due to a perL1-associated olfactory deficit with fertilized females, since perL1 males were unable to discriminate behaviorally between fertilized and virgin females. 2. Other long-period males, including perL2 males and transgenic perL1 males bearing a truncated form of the per+ gene, were conditioned normally by fertilized females. Thus, the courtship inhibition defect is specific to the perL1 mutant strain. 3. perL1 (and other per mutant) flies showed normal acquisition and retention of a classically conditioned olfactory avoidance response. 4. Results from a new conditioned courtship inhibition experiment are presented; males exposed to fertilized females during training showed further courtship inhibition during subsequent exposure to fertilized females. From the perspective of learning theory, this can be viewed as a savings experiment.

Regulation of tyrosine hydroxylase-containing amacrine cell number in larval frog retina

To determine whether production of new neurons of a particular type is regulated by the presence of previously differentiated neurons of the same type, we ablated all tyrosine hydroxylase immunoreactive (THIR) cells from larval frog retina with the neurotoxin 6-hydroxydopamine, and examined the retinas in subsequent weeks for newly generated THIR neurons. Three weeks after neurotoxin administration, new THIR cells appeared near the zone of neural proliferation at the ciliary margin at a higher density than that of normal retina, while the densities of other amacrine cell types, serotonin (t-HT) immunoreactive and substance P immunoreactive (SPIR), remained the same as controls. Thus the production of new retinal TRIR cells is selectively up-regulated following ablation of previously differentiated cells of this type.