We want to be able to engineer animal behavior. Our approach involves two steps: 1) understand, at the molecular level, how behaviors are generated, and 2) use the knowledge in step one to develop tools to manipulate these behaviors. These two steps are not necessarily sequential. Understanding the molecular basis of behavior has allowed us to genetically engineer reagents that allow us to do experiments to better understand behavior.
Because of its ease of manipulation and powerful genetics, we have been concentrating our efforts on the nematode (round worm) Caenorhabditis elegans. C. elegans has a simple and invariant nervous system of 302 neurons whose anatomy and connectivity has been completely described. It exhibits a wide range of behaviors including: eating, egg-laying, mating, chemo- and thermotaxis. With the completion of the worm genome sequence, it has become clear that most of the major gene families of vertebrates are also represented in C. elegans. Thus, molecular mechanisms underlying C. elegans behavior are bound to be of general relevance.
The role of M3 neuron of the C. elegans pharynx
Our work has concentrated on the eating behavior of C. elegans. Worms have a neuromuscular pump called the pharynx that sucks food (bacteria) out of the environment, grinds it up into a digestible pulp, and deposits it in the intestine. The pharynx consists of twenty muscles and twenty neurons. The pharyngeal muscle pumps rhythmically (much like the vertebrate heart muscle), all of the muscles contracting simultaneously. This is advantageous because it allows us to study the electrical activity of the pharyngeal muscle using a technique, similar to the electrocardiogram, that we call an electropharyngeogram (EPG for short). By studying the EPG we found that the duration of the muscle action potential during a pump is regulated by a fast inhibitory pharyngeal motor neuron called M3 that uses glutamate as its neurotransmitter.
To better understand the molecular basis of neurotransmission, we have combined genetics and electrophysiology to characterize mutants that affect neurotransmission by the M3 neuron. We showed that of of the genes that is necessary for M3 neurotransmission, eat-4, acts presynaptically (Dent et al., 1997) in M3. eat-4 turns out to encode a vesicular glutamate transporter and appears to be necessary for neurotransmitter release in all neurons that release glutamate as their neurotransmitter.
Another gene that is necessary for M3 neurotransmission, and which has formed the basis of much of our work, is avr-15. avr-15 encodes a subunit of an glutamate-gated chloride channel (GluCl) (Dent et al., 1997). These channels are underlie an unusual type of neurotransmission, fast inhibitory glutamatergic neurotransmission, that appears to be specific to invertebrates. In the case of C. elegans, GluCls are the receptors in the pharyngeal muscle that allow the muscle to respond to glutamatergic neurotransmission by the M3 neuron. Though specific to invertebrates, these inhibitory glutamatergic ion channels are very similar in structure to the vertebrate glycine-gated chloride channel.
The genetics of ivermectin resistance
As it turns out, the glutamate-gated chloride channels are also the receptors for the antiparasitic drug ivermectin. Ivermectin is widely used in both veterinary and human medicine. Most people are familiar with ivermectin because it is the active ingredient in HeartGuard, a treatment for heart-worm (Dirofilaria immitis) in dogs. However, ivermectin is also part of the World Health Organization's efforts to eradicate river-blindness, a disease caused by the nematode Onchocerca volvulus . River- blindness, which affects some 20 million people in sub-Saharan Africa, cause blindess in ~30% of its victims if untreated.
We and others have shown that ivermectin activates GluCls. Since GluCls are inhibitory chloride channels, ivermectin has the effect of inhibiting the activity of excitable cells such as muscles and neurons that express GluCls. In particular, we have shown that ivermectin inhibits pharyngeal muscle contraction by activating the avr-15-encoded GluCl. This prevents worms from eating and they eventually starve.
Ivermectin affects other GluCls as well and this has important implications for the use of ivermectin and for the design of antiparasitic drugs. We have characterized mutations in two other genes encoding subunits of GluCls, gcl-1 and avr-14. Interestingly, a homozygous mutation in any one of these channels does not make worms resistant to ivermectin. The only double mutant that shows resistance is the avr-14; avr-15 double and the resistance exhibited by these worms is only a modest ~10-fold relative to wild type. In contrast, the avr-14; avr-15 gcl-1 triple mutant is >4,000 fold more resistant than wild type. Thus, each of the genes encoding GluCl subunits independently makes worms sensitive to ivermectin and therefore all three must be mutated to confer high level resistance (Dent et al. 2000). Surprisingly, the triple mutant worm displays only relatively subtle behavioral phenotypes. Our results suggest that: 1) because of the low probability that the worms will acquire multiple mutations in GluCl genes, resistance may be slow to evolve and 2) monitoring the presence of mutations in the GluCls of parasitic worms may help prevent the spread of ivermectin resistance before it becomes a problem.
Problems We'd Like to Solve
Mechanism of Ivermectin Action on Channels
We now know GluCls are the targets of ivermectin, but how does ivermectin make the GluCl channels open? We are beginning to address this question by looking for amino acids that make GluCl channels sensitive to ivermectin.
What role do the GluCls play in behavior
Do worms get Attention Deficit Disorder? The GluCls encoded by avr-14, avr-15 and gcl-1 are necessary for behaviors that, though subtle, are nevertheless beneficial to the survival of the worm in the wild. One of the behavioral defects found in the avr-14 mutants is a sort of attention deficit disorder; where a wild type worm would crawl in a straight line, avr-14 worms back up and turn around repeatedly. We are very interested in role that the avr-14-encoded GluCl plays in the neuronal circuit that guides the decision to either keep crawling or turn around.
Why so Many Channel Subunits?
The GluCls belong to the superfamily of ligand-gated ion channels. Like other members of this superfamily, the GluCl subunits are encoded by multiple genes. And like other superfamily members, the subunits encoded by GluCl genes have similar properties; avr-14, avr-15 and gcl-1 are ~80% identical at the amino acid level and all three form homomeric channels that respond strongly to ivermectin and weakly to glutamate. Why do organisms need so many genes encoding similar proteins? Promoter fusion constructs with green fluorescent protein (GFP) indicate that avr-14 and avr-15 are expressed in mostly non-overlapping patterns. Thus, two hypotheses could explain the presence of so many similar channels. Hypothesis #1, the genes produce functionally interchangeable proteins but, for some reason, three genes, each with different, relatively simple promoter/ enhancers elements evolved rather than one gene with a complex promoter/enhancer element. Hypothesis #2, each of the subunits could have subtle differences that make them uniquely suited to function at the particular synapse where they are found in vivo.
To address this question, one would like to swap promoter and coding regions for the various genes to see if the coding regions are functionally interchangeable. Employing this approach, we have shown that if you express avr-15, avr-14 or gcl-1 coding regions in the pharyngeal muscle of an avr-15 mutant, only the avr-15-encoded GluCl subunit rescues M3 neurotransmission. Thus avr-15, which normally mediates M3 neurotransmission, is uniquely suited to function at the M3 synapse. These results refute hypothesis #1. The question now is, what are the important differences in these channels? We are addressing this question by making chimeric channels and testing their ability to rescue M3 neurotransmission.
Using GluCls to control behavior
To find out what specific subsets of neurons do in a free-ranging intact animal, it would be useful to have a generally applicable means of turning off those small subsets of neurons and looking at the effect of the on behavior.
In theory, any cell that expresses a GluCl with be turned off by ivermectin. Thus, starting with an ivermectin resistant strain of C. elegans, we should be able generate a transgenic worm with a GluCl expressed in any neuron or muscle for which we have a specific promoter. The expression of the GluCl channel will not affect the worm until ivermectin is added. When ivermectin is added, only the cells expressing the GluCl will be affected. Thus we have a specific means of turning off a neuron and looking at the effect on behavior.
We have demonstrated the feasibility of this approach in worms using two different promoters. When the avr-15-encoded GluCl subunit was expressed in the body wall muscles, which drive locomotion, exposure to ivermectin paralyzed the worms. When the avr-15 GluCl subunit was expressed in the mechanosensory neurons that mediate the sensitivity of worms to light touch, exposure to ivermectin made them insensitive to touch. Our current goal is to extend this technique to mammals.