Lab Discoveries
Major Discoveries (most recent first)
Integrin function is required for working memory in the mouse
We are interested in understanding the roles of integrins in synaptic plasticity and memory formation in vertebrates. Using conditional knockout mice, Chi-shing Chan and co-workers have demonstrated that mutants lacking β1-integrins have reduced basal synaptic transmission and are defective in long-term potentiation in the hippocampal Schaffer collateral synapses, while the corresponding fiber volley and paired-pulse facilitation remain normal. These data suggest that β1 deficiency results in a postsynaptic deficit in excitatory synaptic transmission. Interestingly, these β1 knockouts are also impaired in a T-maze non-matching-to-place working memory task. The latter discovery is exceptionally important, since the working memory deficit is very specific and little is known about the molecular and cellular underpinnings of working memory. We are currently trying to establish new assays for working memory and study the functions of β1 and other integrin subunits during working memory formation. These results were reported as Chan et al. in Journal of Neuroscience in 2006.
A delayed memory trace in the DPM neurons
We have extended our studies of elucidating the nature of memory traces in the Drosophila brain by function optical imaging. Most recently, Dinghui Yu discovered that DPM neurons form a delayed memory trace. The pairing of an odor with electric-shock was shown to increase the odor-evoked calcium signals and synaptic release from DPM when tested after conditioning. Most interestingly, these memory traces form in only one of the two branches of the main DPM neuron process, so that the memory traces form in a branch specific manner. Moreover, memory trace formation requires the expression of the wild-type amnesiac gene in the DPM neurons. The cellular memory traces formed in DPM neurons first appear at 30 min after conditioning and persist for at least 1 hr, a time window during which DPM neuron synaptic transmission is required for normal memory. These results were published as Yu et al. in Cell 2005.
Memory trace formation by synaptic recruitment
Distinct odors are encoded by the nervous system in defined patterns of neuronal activity. In the olfactory bulb of vertebrates or the homologous antennal lobe of insects, odor quality is represented by stereotyped patterns of neuronal activity that are reproducible within and between individuals. Using optical imaging to monitor synaptic activity in the Drosophila antennal lobe, Dinghui Yu demonstrated that classical conditioning rapidly alters the neural code representing the learned odor by recruiting new synapses into that code. After pairing an odor conditioned stimulus with an electric shock unconditioned stimulus, new projection neuron synapses are recruited into the brain's representation of the learned odor. In other words, synapses that were not activated by the odor prior to conditioning are activated by the odor after conditioning. Dinghui also found that different odors recruit different groups of projection neurons into the spatial code. The cellular events underlying this form of memory trace — synaptic recruitment — are most likely occurring within the projection neurons themselves, since similar events were not observed in neurons known to be presynaptic to projection neurons. These results were published as Yu et al. in Neuron in 2004.
TARGET to regulate Drosophila transgenes in both time and space
Sean McGuire, a student in the laboratory, developed the TARGET system to regulate the temporal expression of transgenes. The system employs the classical GAL4:UAS approach for spatially-restricted transgene expression, and adds on a temporal feature through the conditional repression of GAL4 by a transgene expressing a temperature-sensitive repressor named GAL80ts. The results of using TARGET to rescue the learning defect associated with the rutabaga mutant were published as McGuire et al. in Science in 2003.
The synaptic output of mushroom body neurons is required for odor memory retrieval but not acquisition or consolidation
Schematic diagram of the effect of Shibire on synaptic transmission
view larger image
image courtesy of Ron Davis Laboratory
Sean McGuire also pioneered the use of a UAS-Shibire[ts] transgene to conditionally inactivate synaptic transmission from the mushroom bodies as the animals learn about odors. The results were published in Science as McGuire et al. in 2001 and demonstrated that synaptic output from mushroom bodies is not required for acquisition or consolidation, but only for retrieval.
Gene-Switch to regulate Drosophila transgenes in both time and space
Lin Zong, in collaboration with Gregg Roman, developed a hybrid transcription factor transgene known as Gene-Switch (Roman et al, 2001) that allows for the conditional expression of a UAS-transgene in both time and space. The transcription factor contains the GAL4 DNA binding domain, an activation domain, and the ligand-binding domain from the human progesterone receptor. Vectors (Roman and Davis, 2002) were developed that use enhancer detection for driving Gene-Switch, or that allow the use of defined promoters or enhancers to drive Gene-Switch. Zhenmei Mao used Gene-Switch to show that rutabaga function only in the mushroom bodies, and only during adulthood, is sufficient to rescue the memory defect of rutabaga mutants. These results were published as Mao et al. in PNAS in 2004.
fasII is expressed preferentially in mushroom bodies and is required for olfactory learning
Expression of fasII in the mushroom bodies
image courtesy of Ron Davis Laboratory
An enhancer detector line was also recovered at the fasII locus, and Yuzhong Cheng used this to demonstrate that this cell adhesion molecule is required for normal olfactory learning. The fasII protein is highly expressed in mushroom body neurons. Furthermore, Yuzhong provided behavioral evidence to support the assignment of fasII in the acquisition of olfactory memories, rather than in consolidation or retrieval. Results were published in Cell as Cheng et al. in 2001.
Mushroom bodies contain three classes of neurons
Three classes of mushroom body neurons
image courtesy of Ron Davis Laboratory
Jill Crittenden, a student in the laboratory, deduced from an extensive series of immunohistochemistry experiments that mushroom bodies consist of three classes, which we named α/β, α'/β', and γ neurons. This important neuroanatomical discovery in 1998 by Crittenden et al. formed the foundation for experiments directed towards understanding the specific role of each class of mushroom body neurons.
Volado encodes a novel alpha integrin
A new learning mutant, also isolated from our enhancer detector screen, proved to encode an integrin. This was important because it was the first demonstration that integrins are involved in behavioral plasticity. The Volado-encoded integrin was also shown to be expressed preferentially in mushroom body neurons. These studies were reported as an article in the journal Nature as Groteweil et al. in 1998.
roX1 specifies a non-coding RNA involved in dosage compensation
roX1 RNA binding to the X-chromosome in salivary gland nuclei
image courtesy of Ron Davis Laboratory
Vicky Meller led our studies of an enhancer detector insertion that expressed reporter in the mushroom body neurons of females but not males. Interestingly, the element marked the location of a gene expressed only in male animals and not females, and that had no open reading frames, indicating that the product was a non-coding RNA. Furthermore, this RNA, later known as roX1 (for RNA on the X) was found to make up chromatin of the X chromosome but not autosomes. roX1 RNA is part of the dosage compensation system of Drosophila. Results were published in Cell as Meller et al. in 1997.
Transgenic rescue of the dunce learning phenotype
Brigitte Dauwalder, a postdoc in the laboratory, showed that a normal dunce transgene expressed with the heat shock promoter was capable of rescuing the behavioral defect of dunce mutants. She also demonstrated, ironically, that expression of the rat homolog of dunce rescues the mutant better than the Drosophila transgene. Results were published in the Journal of Neuroscience as Dauwalder and Davis in 1995.
Protein kinase Acatalytic subunit is expressed at high levels in the mushroom bodies and is required for normal learning
Makis Skoulakis, a postdoc in the laboratory, discovered that one of our enhancer detector lines disrupted the gene for protein kinase A catalytic subunit. In addition, this protein, like dunce and rutabaga, is also preferentially expressed in mushroom bodies, cementing the relationship between cyclic AMP signaling within mushroom bodies for normal learning. Results were published in Neuron as Skoulakis et al. in 1993.
Rutabaga encodes adenylyl cyclase and is expressed at high levels in mushroom body neurons
Enhancer detector line of rutabaga highlighting the mushroom body neurons
image courtesy of Ron Davis Laboratory
Pyung-Lim Han, a graduate student in the laboratory, found from the enhancer detector screen that several of the new lines mapped to the genetic position of rutabaga. He cloned the insertions, and found that they were within an adenylate cyclase gene, and that they disrupted its expression. These experiments were collaborative with Lonny Levin and Randy Reed. He further found that the rutabaga gene was highly expressed in mushroom bodies, much like our previous observation for dunce. These observations solidified that cyclic AMP signaling is intimately connected to learning, and that mushroom bodies are a primary site for olfactory memory formation. Results were published in Neuron as Han et al. and in Cell as Levin et al. in 1992.
Enhancer detector screening for other learning and memory mutants
Han pursued a vigorous histological screen using enhancer detection and searched for genes other than dunce that are expressed preferentially in the mushroom body neurons. Many new lines were isolated and some subsequently shown to exhibit memory deficits. Indeed, this screen led to the molecular isolation of an additional 5 learning mutants, including rutababa, DCO, leonardo, Volado, and fasII. Results were published in the Journal of Neurobiology as Han et al. in 1996.
Mushroom bodies as a neuroanatomical center for olfactory memories
Alan Nighorn, a graduate student in the laboratory, made a fundamental observation published in Neuron as Nighorn et al. in 1991 that changed the direction of the laboratory and the field of Drosophila learning and memory. He found by immunohistochemical studies as well as RNA in situ hybridization that the dunce gene is expressed quite preferentially in Drosophila mushroom body neurons. This brought a cellular focus to our studies as well as those of other laboratories since until this point there had been no connection between genes involved in learning, their products, and specific neurons in the brain. The observation that dunce was beautifully expressed in the mushroom bodies brought a focus on these neurons as a site for memory formation.
Cloning of the bovine and human genes for the cGMP phosphodiesterase involved in visual transduction
Steve Pittler, a graduate student in the laboratory, used the lab’s knowledge of cyclic nucleotide phosphodiesterases to clone and characterize the alpha subunit of the bovine and human cyclic GMP phosphodiesterase that is used in photoreceptor neurons for visual transduction. Results were published in Genomics as Pittler et al. in 1990.
The dunce locus is among the most complex of Drosophila genes, and includes numerous other genes within its introns
In a 1987 Nature article published as Chen et al., Chun-nan Chen showed that the dunce locus is extremely complex, with an intron of 79 kb that contained several other unrelated genes. This report of a nested gene arrangement was the second of its kind; the first example was the GART locus that was characterized by Steve Henikoff. Graduate student Yuhong Qiu further characterized the gene and demonstrated that the gene is at least 148 kb in genomic length, extends across 9 salivary chromosome bands, and encodes at least 10 different RNAs through alternative splicing, the use of least 5 different transcription start sites, and the use of multiple polyadenylation sites. Results were also published in the Journal of Neuroscience as Qiu et al. in 1991 and in Genes and Development as Qiu and Davis in 1993.
The mammalian homologs of dunce are the target for rolipram, a drug with antidepressant activity
We noticed from the literature in 1988 that a promising new antidepressant, named rolipram, was an effective inhibitor of cyclic nucleotide phosphodiesterase in crude homogenates of mammalian cells. However, this issue was complex since there exist several dozen cyclic nucleotides phosphodiesterase isozymes in mammalian cells. We isolated the first clones for the rat and human homologs of dunce and then used our expression systems to show that the encoded phosphodiesterases are indeed targeted by rolipram. This was a major observation that formed the first connection between Drosophila learning mutants and mammalian behavior. Mutation at the dunce locus in Drosophila produced a learning defect, whereas its pharmacological inhibition in humans produced an elevation of mood. These studies were published as Henkel-Tiggs and Davis, 1990, and Livi et al., 1990.
The dunce encodes a cAMP specific form of phosphodiesterase activity
The fact that dunce mutations decreased cAMP phosphodiesterase activity and elevated cAMP levels produced the alternative hypotheses that dunce was either the structural gene for cAMP phosphodiesterase, or that it encoded a regulatory molecule for cAMP phosphodiesterase. Graduate student Chun-nan Chen demonstrated that the dunce protein products have a very high sequence homology with other phosphodiesterases, and later, Yuhong Qiu, anonther graduate student in the laboratory, showed that a dunce cDNA could be expressed in yeast to produce cAMP phosphodiesterase activity. These results established unambiguously that dunce does encode cAMP phosphodiesterase. These studies were published as Chen et al., 1986, and Qiu et al., 1991.
Restriction site polymorphisms for mutant mapping in Drosophila
In 1984, Ron Davis, along with Norman Davidson, published the first account of using restriction site polymorphisms as genetic markers in Drosophila. Results were published in Molecular and Cellular Biology as Davis and Davidson in 1984.
The learning mutant dunce, disrupts cAMP phosphodiesterase activity
In a series of publications from 1978-1984, Ron Davis characterized the cyclic nucleotide phosphodiesterases in Drosophila and demonstrated that dunce mutations affected the activity of a cAMP-specific phosphodiesterase and caused a dramatic elevation in cAMP levels. The initial studies on dunce were published along with Duncan Byers and John Kiger in 1981 in the journal Nature as Byers et al. This became the first biochemical connection associated with a Drosophila learning mutant and initiated the accumulation of knowledge about the association of cAMP signaling and memory formation. Results were also published in the Journal of Cell Biology as Davis and Kiger in 1981.