Courtesy of Genevieve Anderson

If not for Nobel laureates Thomas Hunt Morgan, Eric R. Kandel, and Sydney Brenner, the notion of a general behavioral model might seem odd. Behaviors, after all, are determined by an animal's evolutionary history and ecological niche. They are often idiosyncratic, shared in detail only by closely related species.

But, thanks to Morgan's research in the early 20th century, and Kandel's and Brenner's work over the past 35 years, the fly Drosophila melanogaster, the mollusk Aplysia californica, and the worm Caenorhabditis elegans have become general behavioral models. The newest member of the club is the mouse.

This quartet yields broadly applicable behavioral findings for two reasons: First, these animals are unusually amenable to cellular and molecular experimentation; second, such experimentation has turned up certain genes, proteins, and cells that underlie behavior across many species. Evolution did not "completely reinvent the wheel and come up with a new set of molecular rules for each phylum," notes Aplysia expert Thomas J. Carew, at the University of California, Irvine.

LEARNING FROM SEA SLUGS Lacking a cortex, Aplysia has just 20,000 neurons clustered into ganglia. As such, its nervous system appears incommensurable with those of higher organisms. But this hand-length, maroon sea slug has one quality trumping interspecies differences: huge neurons, many with cell bodies hundreds of microns across. Biologists can easily image and manipulate these neurons to determine their firing properties and responses to stimulation. Popular research areas include the mollusk's feeding behavior and its withdrawal reflexes when it is touched.

One important discovery, says Carew, is that facilitation--a phenomenon involving enhanced neurotransmitter release into the synapses separating neurons--underlies a simple form of learning known as sensitization. Other findings have elucidated the signal-transduction pathways triggered by learning. Cyclic AMP (cAMP) activates cytoplasmic kinases (e.g., protein kinase A), which translocate to the nucleus where they activate transcription factors (e.g., cAMP response element-binding protein). These factors then turn on genes whose protein products cause long-term changes in the neuron.

Many findings in Aplysia have been replicated in Drosophila and mice. Experiments on transgenic and knockout mice, for example, show that synaptic plasticity relies on several kinases and transcription factors first explored in Aplysia. Conversely, knowledge gleaned from higher organisms might apply to sea slugs. Using a paradigm tested in humans, Carew learned that training the mollusk induces long-term memory if sessions are separated in time but not if they are massed together.

Aplysia has two major limitations as a model for higher organisms: a modest behavioral repertoire, and a genome that has not been sequenced (unlike the genomes of the other three behavioral models). To manipulate this mollusk genetically, researchers inject mRNA directly into its neurons.

LORDS OF THE FLIES Genetic plasticity is Drosophila's chief advantage. In 1915, the Columbia lab of fruit-fly pioneer Morgan conducted the first behavioral genetics study of any organism, recounts Brandeis University biologist Jeffrey C. Hall. Since then, scientists have discovered or created thousands of fly mutants. Tools for investigating Drosophila include heat-inducible and tetracycline-regulated transgenes, transposable P-elements, and the GAL4-UAS system, which allows precise spatial control of transgene expression.

Many fly researchers are examining circadian cycling between activity and inactivity, as well as learning and memory.¹ Some are focusing on courtship, geotaxis, and reactions to odor and taste. At least 15 homologs of fly genes implicated in these behaviors have been found in other species, Hall says.

Drosophila learning occurs during natural behaviors--courtship, for example, is not totally hard-wired--and in conditioning experiments. Martin Heisenberg, at the University of Würzburg in Germany, has trained flies to avoid the heated side of a chamber and devised a complex flight simulator to test visual learning. One popular type of apparatus shocks flies as they sniff an odor. Some mutant or transgenic flies later forget to avoid the odor.

Neuroscientist Jerry C.P. Yin, at Cold Spring Harbor Laboratory, learned that one murine transgene actually enhances certain forms of Drosophila memory. It encodes a specific form of the signal-transduction enzyme protein kinase C. Based on this protein's functions in other cell types and species, Yin speculates that it allows a neuron to tag its recently active synapses.

Two other Cold Spring Harbor neuroscientists, Tim Tully and Josh Dubnau, used the odor/shock assay to uncover about 60 putative memory genes. Switching to DNA chip technology, Tully and Dubnau identified 42 genes that turn on or off during memory formation. Prominent among the genes implicated by both methods were staufen, whose protein product appears to be involved in transporting mRNAs to synapses, and pumilio, whose product seems to help repress translation during mRNA transport.

Tammy Irvine, Rear View Illustration

Hall specializes in the fruitless gene, which encodes a transcription factor critical to male courtship. He regards fruitless as the best example of a single gene specifying a set of behaviors. His lab found that fruitless is expressed throughout the fly's nervous system--a discovery, he notes, that is consistent with the gene's broad effects.

Genetic malleability, Drosophila's greatest strength, also can be a weakness. In rapidly breeding mutant populations, further mutations often cause loss of phenotype, which is difficult to reverse. (Labs cannot preserve the original phenotype because there are no reliable methods to freeze and thaw fly embryos.) For neurobiologists, the fly's main drawback is that its 150,000 or so neurons are too small to manipulate in situ, except at the neuromuscular junction.

NOTHING TO HIDE The nematode C. elegans joined the behavioral-model menagerie basically because it has nothing to hide. The tiny transparent worm's nervous system has been completely plotted, revealing 302 neurons and 5000 synapses. With this knowledge in hand, researchers employ various cell-ablation and gene-manipulation techniques to link behaviors to specific cells and genes. Investigators also can record from nematode neurons and recently began culturing them.

Worms have a limited behavioral repertoire. Besides life- and species-preserving activities such as feeding, mating, and egg-
laying, nematodes "can't do much--swim forward, swim backward, stop, start, curl in a circle," observes psychology professor Catherine H. Rankin, at the University of British Columbia.² Nevertheless, their exquisite sensitivities to temperature and ambient chemicals facilitate behavioral conditioning experiments. Homologs of Drosophila learning-associated genes are known, but their effects on C. elegans learning are unclear, says Rankin. Altering these signal-transduction genes, she explains, often results in an uncoordinated worm, a common and unrevealing mutant phenotype.

Even when a mutation's impact seems unambiguous, the full story is probably far from simple. Five years ago, neuroscientists Mario de Bono, now at the Medical Research Council's Laboratory of Molecular Biology in Cambridge, England, and Cornelia I. Bargmann at UC-San Francisco, discovered a loss-of-function mutation in a receptor gene that switched C. elegans from a solitary food-forager to a "social" one that gathers with its mates on their common source of nourishment, a lawn of bacteria.

Follow-up work suggests the existence of "multiple layers of antagonistic stimuli that are regulating whether you are social or solitary," says de Bono. "There's often this assumption that because a single gene flips behavior from one form to another, it is the critical gene." But, he demurs: "You can still have that effect when you have a gene that's only one player in a larger number of players."

Tammy Irvine, Rear View Illustration

KNOCKOUTS AND THEIR DISCONTENTS Rats were long the rodent behavioral model of choice because of their intelligence and large brains. Over the past decade, however, transgenic mice have increasingly hogged the spotlight. (Transgenic rats are relatively rare, because foreign DNA does not incorporate readily into the genomic DNA of rat oocytes, explains Markus Heilig, at the Karolinska Institute, Sweden.) Jacqueline N. Crawley, chief of the National Institute of Mental Health's behavioral genomics section, notes that transgenic mice have become invaluable in experiments involving learning and memory, motor disorders such as Parkinson disease, and obesity.³

Mice lacking certain genes can display complex behavioral phenotypes, such as social deficits (knockout of the hormone oxytocin); less huddling and nest-building (knockout of the intracellular signaling molecule dishevelled-1); and male aggression (knockout of neuronal nitric oxide synthase). But Crawley cautions that different murine strains with an identical genetic alteration might each exhibit a unique phenotype. The likely reason: Each strain harbors a different set of polymorphisms that temper the genetic alteration's effect.

Transgenic studies involve other complexities and pitfalls. Crawley observes, for example, that some murine wild-type strains are already so aggressive that an aggression-causing mutation might be undetectable. She also warns of "a lot of variability in behavior that requires larger numbers of animals and more rigorous statistics than molecular geneticists are used to." When these requirements are not met, she adds, investigators often overinterpret results, causing the field of behavioral neuroscience to lose some credibility.

Crawley's mantra is that "behavior is not so simple." But that's a challenge that Drosophila expert Hall relishes. He insists that biologists "should want the phenomenon [they study] to be complicated, because life is complicated."

Douglas Steinberg (dougste@attglobal.net) is a freelance writer in New York City.