Title: Specific roles for DEG/ENaC and TRP channels in touch and thermosensation in C. elegans nociceptors
Here they were interested in discovering which proteins mediate nociception - the sensation of pain. Neurons that detect pain are referred to as nociceptors, and are thought to be multimodal - that is, they respond to many kinds of sensory inputs that generate pain, like heat, cold, harsh touch, extreme pH, or noxious chemicals. This paper seeks to answer the question of how nociceptors obtain their multimodality. Do they have single receptors that respond to all these inputs? Or do they have many receptors? At which point do these sensory inputs converge onto the same signal transduction path?
As for 90% of the papers I've read recently, they work with C. elegans, and in particular focus on the PVD neuron, which has been shown to mediate the response to harsh touch (based on previous laser ablation work).
For some background, PVD does not look like the typical C. elegans neuron. Rather than a single more or less unbranched process, PVD has a legitimate dendritic arbor that spans basically the entire worm body wall. Check out the pictures from wormatlas. It's gorgeous!
They begin by observing the calcium response of PVD to harsh touch, and confirm that the neuron does indeed show a calcium influx in response to harsh touch (here, this means touching the worm with platinum wire or glass needle). Note that this occurred while the worms were fixed to a surface with cyanoacrylate glue, and the glass needle was "driven into the worm ~30-50um at speed of 2.8um/s. Stimulus duration was ~50ms." This seems contradictory, since moving 30-50um at 2.8um/s would take 10s of seconds, not 50 ms! I must be missing something.
Unfortunately, they don't discuss where the calcium response is, or where the harsh touch was applied! This neuron extends processes everywhere, so is the calcium influx uniform throughout? Or localized? What was plotted? Total or only somatic dynamics?
After verifying the PVD yields a calcium response to harsh touch but not gentle touch (same thing but only driven 10um into the worm), they investigated other noxious stimuli that PVD might respond to. They did this by ablating PVD and testing the worms response to various noxious stimuli (what stimuli these were aren't mentioned), and found that PVD-ablated worms don't do omega turns when exposed to cold. I presume that in wild type worms this is a thermotactic response ("it's getting colder so I must be going the wrong way - I should turn around!"). Further measurements with calcium imaging in PVD showed that PVD responded to 5-degree C decreases in temperature, but not smaller temperature decreases, and not temperature increases. They thus conclude that PVD is basically a bimodal nociceptor, responding to harsh touch and cold shock.
How does PVD sense these stimuli? They start with harsh touch, and first investigate whether MEC-10 is necessary. MEC-10 is a mechanosensory DEG/ENaC (the latter meaning epithelial sodium channel) which is known to be needed to respond to gentle touch in the ALM/AVM/PLM neurons. mec-10 mutants don't generate a PVD calcium respond to harsh touch, and using two semi-specific PVD promoters (each of which also drives mec-10 expression in one other neuron type), they show that they can rescue the PVD calcium respond with PVD-specific mec-10 expression.
As an aside, it's worth note that they're good enough to show the animal-to-animal variability in calcium responses, which are huge in all the backgrounds that do respond. Some animals don't respond, some animals yield 60-70% fluorescence changes, and everywhere in the middle. I still find this striking, how variable these responses are, and why they might be this way. It could be that their measurement system just introduces a lot of noise, but I also wonder how much of this is simply due to the ongoing dynamics in the network at the time the stimulus is applied.
Now back to the paper. Aside from mec-10, they were interested in other genes that might be essential for PVD sensing harsh touch. Based on transcriptional profiling of PVD, they found three other DEG/ENaC channel genes (aside from mec-10) that were highly expressed in PVD, which they reasoned might be involved in sensation of harsh touch. Deletion mutants of two of them showed normal harsh touch response behavior and calcium transients, so they rule them out. The other one, F25D.14, didnt have a deletion mutant, so they knocked it down with RNAi, and found that it was deficient in harsh-touch calcium transients. Interestingly, they note that its escape behavior was messed up in a mec-4 background, but don't comment on the effect of the RNAi on behavior in wild type background. I'm a little confused about this, because this isn't stated in the figure (the figure clearly states that these experiments were done in wild-type animals).
Regardless, F25D1.4 appears to be necessary for harsh touch, and they name this gene degt-1.
Using GFP and RFP fusions, they show that DEGT-1 and MEC-10 colocalize in puncta along the dendrite, and also colabel the nucleus (though not in a punctate manner). In a mec-10 null background, DEGT-1::mCherry-RFP no longer localizes to puncta, but instead is diffuse throughout the entire process. This would be consistent with MEC-10 interacting with DEGT-1 as part of a mechanosensory complex.
They then go on to show that degt-1 is expressed in ALM (a 'body touch' or 'gentle touch' mechansensory neuron). ALM responds to both gentle and harsh touch, but in a mec-4 background, ALM only responds to harsh touch. From previous studies, it was known that mec-10 is also necessary for gentle touch, and thus mec-10 seems to be involved in both gentle and harsh touch sensation. degt-1 RNAi affected harsh touch responses, but not both. They thus predict that mec-10 might form complexes with mec-4 to mediate soft touch, and complexes with degt-1 to mediate harsh touch. To test this, they overexpress degt-1, predicting it will thus be able to outcompete mec-4 for mec-10 partners. They find that gentle touch responses (calcium and behavior) are compromised, consistent with this hypothesis.
Now they move onto exploring how PVD senses cold shock. mec-10 mutants and degt-1 RNAi don't seem to change the PVD calcium response to cold shock, so it's clearly mediated by a different mechanism. THey next investigated two TRP channels expressed in PVD - TRPA-1 and OSM-9. Of these, trpa-1 mutants have no cold-shock calcium response, whereas osm-9 do. Using a GFP fusion, they show TRPA-1 is localized to the soma.
They next show that a mouse TRPA-1 also confers cold-sensitivity when expressed in PVD in a trpa-1 mutant, suggesting that it functions as a cold-responsive sensor in mouse nociceptors. They also show that trpa-1 expression in the normally heat-sensitive but cold-insensitive FLP neurons and in the touch-sensitive but temperature-insensitive ALM neurons is sufficient to obtain calcium responses to cold shock in those neurons.
Finally, they show in TRPA-1 transfected HEK293T cells that the membrane resistance drops drastically in response to cold (15C!). Strangely, they do not show the negative control of untransfected HEK-293T cells... and 15 C is pretty extreme for mammalian cells, so I'm not sure how much I really trust this experiment. Nevertheless, it's pretty clear that trpa-1 is a key cold-sensor in C. elegans.
Questions left over:
- Where is the calcium response in PVD? Is it localized to a nearby section of the dendritic arbor? Or is it everywhere throughout the neuron?
- What noxious stimuli does PVD not respond to? It sounds like they assayed a number of stimuli, but then seemingly don't report the negative results.