Evolution of a polymodal sensory response network
© Srinivasan et al. 2008
Received: 26 September 2008
Accepted: 15 December 2008
Published: 15 December 2008
Skip to main content
© Srinivasan et al. 2008
Received: 26 September 2008
Accepted: 15 December 2008
Published: 15 December 2008
Avoidance of noxious stimuli is essential for the survival of an animal in its natural habitat. Some avoidance responses require polymodal sensory neurons, which sense a range of diverse stimuli, whereas other stimuli require a unimodal sensory neuron, which senses a single stimulus. Polymodality might have evolved to help animals quickly detect and respond to diverse noxious stimuli. Nematodes inhabit diverse habitats and most nematode nervous systems are composed of a small number of neurons, despite a wide assortment in nematode sizes. Given this observation, we speculated that cellular contribution to stereotyped avoidance behaviors would also be conserved between nematode species. The ASH neuron mediates avoidance of three classes of noxious stimuli in Caenorhabditis elegans. Two species of parasitic nematodes also utilize the ASH neuron to avoid certain stimuli. We wanted to extend our knowledge of avoidance behaviors by comparing multiple stimuli in a set of free-living nematode species.
We used comparative behavioral analysis and laser microsurgery to examine three avoidance behaviors in six diverse species of free-living nematodes. We found that all species tested exhibit avoidance of chemo-, mechano- and osmosensory stimuli. In C. elegans, the bilaterally symmetric polymodal ASH neurons detect all three classes of repellant. We identified the putative ASH neurons in different nematode species by their anatomical positions and showed that in all six species ablation of the ASH neurons resulted in an inability to avoid noxious stimuli. However, in the nematode Pristionchus pacificus, the ADL neuron in addition to the ASH neuron contributed to osmosensation. In the species Caenorhabditis sp. 3, only the ASH neuron was required to mediate nose touch avoidance instead of three neurons in C. elegans. These data suggest that different species can increase or decrease the contribution of additional, non-ASH sensory neurons mediating osmosensation and mechanosensation.
The overall conservation of ASH mediated polymodal nociception suggests that it is an ancestral evolutionarily stable feature of sensation. However, the finding that contribution from non-ASH sensory neurons mediates polymodal nociception in some nematode species suggests that even in conserved sensory behaviors, the cellular response network is dynamic over evolutionary time, perhaps shaped by adaptation of each species to its environment.
Despite a more than 1000-fold range in the sizes of different nematode species, nematode nervous systems are composed of a small number of neurons: Caenorhabditis elegans, which is 1 mm long, has 302 neurons , whereas the parasite Ascaris lumbricoides, which grows to be 20 cm long, has 298 neurons . This constancy in number of neurons suggests a constraint in the nervous system of nematodes.
Free-living nematodes use amphids and phasmids as sensory structures to seek food and avoid harmful stimuli [3, 4]. In C. elegans, the functions of several of the amphidial neurons and their roles in various behaviors have been characterized in detail [3–5]. Parasitic nematodes, on the other hand, use amphids to either actively seek the host (Strongyloides stercoralis) [6, 7] or passively seek the host (Haemonchus contortus) [8, 9]. Electron microscopic reconstruction of the amphids of the free-living nematode C. elegans and several parasitic nematodes such as H. contortus and Ancylostoma caninum, have shown that the sensory neuroanatomy is remarkably similar between the different species [1, 6, 7, 10, 11]. For instance, the amphid sensilla in H. contortus, a passively ingested parasite of sheep, shows the presence of 12 sensory neurons just like in C. elegans [1, 8, 9, 11]. This conservation of neuroanatomy between H. contortus and C. elegans is in perfect correlation with the current phylogeny wherein this parasitic nematode is in the same order as C. elegans [12, 13]. However, it is surprising that even in the distantly related nematode parasite S. stercoralis, there is remarkable similarity of the amphid structure with C. elegans [6, 7, 10]. Given this similarity, it is highly likely that similar cells would mediate the same behaviors across these diverse nematode species.
Studies by Schad and colleagues on the role of homologous neurons mediating chemotaxis, thermotaxis and development in parasitic nematodes have shown that positionally homologous neurons perform similar functions in these species, suggesting that neuronal function is conserved between free-living and parasitic nematodes [10, 14, 15]. Infective juveniles in parasitic nematodes use chemical, physical and thermal signals to find their host and resume development [14, 16, 17]. For instance, the finger cell AFD neurons in H. contortus and the dog hookworm A. caninum act as thermosensors just as in C. elegans [17, 18]. Even in a more distantly related nematode S. stercoralis, ablation of the ASE neuron and ASH neuron resulted in loss of attraction to sodium chloride at lower concentrations and avoidance of high salt concentration respectively [19, 20]. These results suggest that C. elegans can serve as a useful model in understanding the role of different sensory neurons in diverse nematode species.
For our comparative analysis, we chose six representative free-living species from different groups of rhabditids with a range of phylogenetic distances to C. elegans (Figure 1B). Based on molecular data, rhabditids can be distinguished into two major clades: the Eurhabditis and the Pleiorhabditis . C. elegans, along with other members of the Caenorhabditis genus, belongs to the Eurhabditis clade . From the genus Caenorhabditis, we chose C. elegans (N2), Ceanorhabditis briggsae (AF16), and the less closely related Caenorhabditis sp. 3 (PS1010) [12, 13]. C. elegans and C. briggsae were isolated from compost and Caenorhabditis sp. 3 (PS1010) was isolated from galleries of palm and sugarcane weevils . Cruznema tripartitum (SB202) is also a member of Eurhabditis but belongs to a different branch than Caenorhabditis . This species has been isolated from different kinds of rotting organic material, and also from garden soil and compost . From the group of diplogastrids, we chose the satellite model system Pristionchus pacificus (PS312) [30, 31], which belongs to a genus that associates with beetles . Panagrellus redivivus (PS2298) was chosen as a representative of the outgroup and has been isolated from sugar-rich environments such as sap of rubber trees and brewery mash  (Figure 1B). The choice of nematode species provided a broad phylogenetic spectrum for our comparative analysis.
Cellular anatomy is essentially invariant between individuals of the same nematode species and the number of neurons is highly conserved between nematode species, allowing the identification of neurons across different species of nematodes [6, 10, 17, 34, 35].
In C. elegans, nose touch response is mediated by two non-amphid neurons, FLP and OLQ, in addition to the ASH neurons . To test whether nose touch response might be mediated by a single sensory neuron (ASH) and to check the role of the non-amphid neurons, we assayed nose touch response in FLP-ablated, OLQ-ablated, FLP/OLQ-ablated and FLP/OLQ/ASH-ablated animals in both C. elegans and Caenorhabditis sp. 3 (Figure 5B). In C. elegans, FLP-ablated animals were significantly less responsive to nose touch compared with unablated animals. OLQ-ablated C. elegans showed no effect on nose touch compared with unablated animals (Figure 5B). However, ablation of both the non-amphid neurons FLP and OLQ in C. elegans resulted in significantly decreased nose touch avoidance compared with either of the single neuron ablations, suggesting that among the non-amphid neurons, FLP plays a major role in mediating nose touch whereas OLQ plays a minor role  (Figure 5B). In C. elegans, FLP/ASH-ablated animals were significantly different from ASH-ablated animals (Figure 5B). Animals lacking FLP, ASH, and OLQ neurons were significantly different from ASH-ablated animals but not different from ASH/FLP-ablated animals (Figure 5B).
In Caenorhabditis sp. 3, the FLP and OLQ neurons have no effect on nose touch response (Figure 5B). Moreover, ablation of both the two non-amphid neurons (FLP, OLQ) did not show any significant difference from either of the single ablated animals (Figure 5B). In Caenorhabditis sp. 3, ablation of all three neurons (ASH, FLP, and OLQ) did not significantly differ from ASH-ablated animals (Figure 5B). These results suggest that the ASH neuron completely mediates nose touch avoidance behavior in Caenorhabditis sp. 3, as compared with three neurons in C. elegans. It is conceivable that in Caenorhabditis sp. 3, other neurons could be involved in mediating nose touch behavior. However, ASH-ablated animals in Caenorhabditis sp. 3 show less of a nose touch response compared with ASH/FLP/OLQ-ablated animals in C. elegans, suggesting that ASH mediates most of the nose touch response in Caenorhabditis sp. 3 (Figure 5B). The sensory response network for mechanical stimuli in Caenorhabditis sp. 3 thus shows a reduction in the number of neurons involved compared with C. elegans.
Comparison of mean response times to 2 M glycerol in free-living nematodes
Mean response time (s)
Ablation of the ASH neuron almost completely abolishes osmotic avoidance response in C. elegans . In the parasitic nematode S. stercoralis, avoidance response to high salt concentrations is completely mediated by the ASH neuron . Ablation of the ASH neuron in the other species resulted in reduced a.i. (Figure 6A). ASH ablation does not completely abolish osmotic avoidance in P. pacificus, suggesting the possibility of ASH-independent mechanisms of osmotic avoidance (C. elegans a.i. = 0.15 vs. P. pacificus a.i. = 0.36; Figure 6A). In C. elegans, ASH, ADL, ASK, and ASE neurons are known to mediate avoidance of chemical repellents . The residual osmotic avoidance seen in ASH ablated animals of P. pacificus could be mediated by these neurons. In C. elegans, ablation of the ADL neuron had no effect on osmotic avoidance (Figure 6B). Ablation of both ASH and ADL neurons in C. elegans did not show a significant change in osmotic avoidance index compared with ASH-ablated animals (Figure 6B). In P. pacificus, ablation of ADL neuron results in a significantly lower avoidance index compared with unablated animals (Figure 6B). Ablation of both ASH and ADL neurons in P. pacificus had a significantly lower avoidance index compared with ASH-ablated or ADL-ablated animals, suggesting an additive effect of both the neurons in osmotic avoidance behavior (ASH/ADL-ablated a.i. = 0.17 vs. ASH-ablated a.i. = 0.36; ASH/ADL-ablated a.i. = 0.17 vs. ADL-ablated a.i. = 0.62) (Figure 6B). Ablation of the ASK and ASE neurons did not change the osmotic sensitivity of P. pacificus, indicating that these neurons do not play any role in osmotic avoidance (data not shown). Hence, we believe that the ASH neuron is the primary sensory neuron for mediating osmotic avoidance in all nematode species, but additional sensory neurons, such as ADL neuron in P. pacificus, may contribute to mediating osmotic avoidance.
In C. elegans, the ADL neuron, along with ASH and ASE neurons, mediates avoidance response to heavy metal ions, but does not play a role on its own in mediating this avoidance . In the dog hookworm A. caninum, both ASH and ADL neurons are required to mediate avoidance of the detergent sodium dodecyl sulphate . Ablation of only the ADL neurons in the hookworm results in an assortment of responses, from avoidance to non-responsive behavior, suggesting that the animal is not able to decide whether to respond to the chemical . In P. pacificus, ablation of ADL neuron resulted in a significant reduction of avoidance response, indicating that this neuron is necessary for mediating avoidance to high osmotic conditions. The role of the ADL neuron during avoidance behaviors in different species thus remains unclear. Perhaps analysis of the genes expressed in the ADL neuron would provide insight into the role of this neuron in the different species.
We cannot rule out the possibility that ablation of specific neurons results in functional replacements by other neurons. However, if this occurs, then this aspect of nematode neurobiology varies among cells, behaviors, and species.
We hypothesized that nematodes would display habitat-specific sensitivity to the same stimuli rather than phylogeny based sensitivity. Therefore, we tested several different concentrations of the different noxious stimuli (chemosensation and osmosensation), and used a cluster algorithm to generate a behavioral dendrogram for the different avoidance behaviors (see Methods).
As with osmotic avoidance behavior, we found that P. redivivus was the least sensitive to glycerol compared with the other species (Figure 7B). This observation correlates with the fact that P. redivivus has been isolated from high osmotic strength environments and hence could be adapted to high osmolarity [33, 44]. However, at the highest concentration tested (4 M glycerol), P. redivivus also exhibited a very high avoidance, as did the other species (Figure 7B). C. briggsae, Caenorhabditis sp. 3, C. tripartitum, and P. pacificus exhibited similar sensitivity to osmotic avoidance (Figure 7B). C. elegans exhibited slightly different osmotic sensitivity different from all these species with the exception of P. redivivus (Figure 7B). Hence, other than P. redivivus, all other species exhibited similar sensitivity to different osmolarity conditions.
Finally, we combined the data of octanol and osmotic avoidance behaviors including the ASH neuron ablation data for all the species and found that P. redivivus exhibited the most different behavior in our analyses (Figure 7C). Based on relative distances computed by the algorithm, C. tripartitum is the next closely related species to P. redivivus. C. elegans, along with the other species, forms a whole branch on the behavioral dendrogram, with P. pacificus being closely related to it. C. briggsae and Caenorhabditis sp. 3 form a sub-branch, suggesting that these two exhibit similar behaviors. Comparing our behavioral dendrogram for octanol and osmotic avoidance behaviors with the phylogenetic tree, we see some interesting features (Figures 7C and 7D). We observe that for these two behaviors, the relative positions of P. redivivus and C. tripartitum resemble that of the phylogenetic tree (Figures 7C and 7D). P. pacificus seems to be behaviorally similar to the branch of Caenorhabditis sp. 3 and C. briggsae. Given the association of Caenorhabditis sp. 3 with rice weevils  and P. pacificus with beetles , this correlation makes sense. Unexpectedly, C. elegans does not exhibit similar behavioral properties like its close siblings C. briggsae and Caenorhabditis sp. 3. These data suggest that sensitivity to different stimuli varies among species and that the differential sensitivity could be linked to the functional sensory receptor repertoire of these species .
Comparison of sensory response networks in free-living nematodes
Caenorhabditis sp. 3
At a cellular level, avoidance response to the chemical 1-octanol was mediated by the ASH neuron in all species. In parasitic nematodes, avoidance of high salt concentration and sodium dodecylsulfate is also mediated by the ASH neuron [20, 21]. However, for nose touch response behavior, we observed a reduction in the number of sensory neurons relative to C. elegans, with only the ASH neuron mediating this response in Caenorhabditis sp. 3 as compared with three neurons in C. elegans (ASH, FLP, and OLQ). On the other hand, we see an increase in the set of sensory neurons mediating osmotic stress with the ADL neuron partially mediating osmotic avoidance in P. pacificus along with the ASH neuron (Table 2). We also observed that sensory sensitivity to certain stimuli varied between the different species tested. These differential responses could be attributed to adaptation of the species to their respective niches. For example, P. redivivus has a slower response time in response to 1-octanol and is highly resistant to high osmotic conditions (Figures 4 and 6). Since P. redivivus was isolated from the sap of rubber trees and brewery mash, these responses could be attributed to adaptation to sugar-rich environments.
The conservation of several properties of a multi-functional neuron across diverse lifestyles is not surprising, given the constancy of neuron number in these nematodes. However, to understand how sensory responses evolve, comparison of closely related species is an essential prerequisite. Such studies can uncover selection pressures that act within a shared evolutionary history to cause either differences in their behavior or changes in cellular contribution to these behaviors . These changes may evolve by alterations in the number and type of sensory receptors each neuron expresses, by the efficacy of signal transduction in sensory neurons, or by the strength of connectivity of the neuron to its postsynaptic partners. Our results provide a clear demonstration of change in the relative contributions of nociceptive neurons in a sensory network in diverse free-living nematode species. These studies allow inference of ways in which sensory responses in free-living nematodes might evolve, and suggest how such evolution might occur in medically and agriculturally relevant nematodes.
All strains were raised at 20°C unless indicated otherwise, using standard methods . Strains used in this study were C. elegans (N2 Bristol isolate) , C. elegans eat-4(ky5) MT6308 , C. elegans glr-1(n2461) KP4 , C. briggsae (AF16), Caenorhabditis sp. 3 (PS1010), P. pacificus (PS312) , C. tripartitum (SB202), and P. redivivus (PS2298/MT8872) .
A stock solution of DiI (2 mg/ml) was prepared in dimethyl formamide and stored at -20°C . Worms from the different species were washed from plates and resuspended in a small eppendorf tube to a 100 μl volume. The worm pellet was washed thrice in M9 buffer and worms allowed to settle down using gravity. To this pellet we added a 1:200 dilution of DiI solution and incubated in the dark for 2–3 hours. After the incubation, the worms were washed three times in M9 buffer and then mounted on a slide to visualize the amphid neurons. For some species, we used a longer incubation time (5–7 hours) to visualize neurons.
For all species tested, we used the L1 larva stage for our ablations as described previously . All behavioral assays were performed as described previously [23, 25, 42]. A detailed description of the ablations and behavioral assays is given in the Supplementary Materials and Methods in Additional file 1. Ablation of one of the two bilaterally symmetrical neurons did not affect the avoidance response to different stimuli.
For octanol avoidance sensitivity, we made different dilutions of 100% octanol (0.1%–100%) and each species was tested with each concentration at least on three different days. Data was represented as average avoidance time for each concentration.
Similarly, for sensitivity to osmotic avoidance, we tested concentrations ranging from 0.1 M to 4 M glycerol. Data was represented as mean a.i. (see Supplementary Materials and Methods in Additional file 1 for details).
For each of the behavioral assays, the normalized a.i. was computed for each species. This was done by normalizing the value for each species at each concentration to the maximum value for that concentration. The normalized values for the different concentrations were then plotted for each species in Matlab. For obtaining the dendrogram, data was then clustered using a Matlab hierarchical clustering algorithm (Euclidean distances and average linkage). The x-axis of the dendrogram indicates relative distance between the different species.
The statistical tests employed are described in Additional file 1.
We thank the Bargmann Lab and the Hart Lab for advice, and Krisha Begalla and Dorota Korta for help with the behavioral assays, Alon Zaslaver for suggesting the cluster analysis and help in generating Figure 7. We thank Karin Kiontke for the C. tripartitum (SB202) strain and advice on phylogeny. Some nematode strains were provided by the Caenorhabditis Genetics Center. We also thank Cori Bargmann, Ryan Baugh, Takao Inoue, Elissa Hallem, Marie-Anne Felix, Ralf J. Sommer and Karin Kiontke for discussions and comments on the manuscript. JS is an Associate and PWS is an Investigator of the Howard Hughes Medical Institute, which supported this research. OD was supported in part by a Summer Undergraduate Fellowship from Caltech.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.