NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Riddle DL, Blumenthal T, Meyer BJ, et al., editors. C. elegans II. 2nd edition. Cold Spring Harbor (NY): Cold Spring Harbor Laboratory Press; 1997.

Cover of C. elegans II

C. elegans II. 2nd edition.

Show details

Section VGenetic and Molecular Analysis of Chemosensation and Thermosensation

A variety of behavioral screens have been used to identify mutant animals with defects in chemosensory or thermosensory behaviors (Table 3) (Ward 1973; Dusenbery et al. 1975; Hedgecock and Russell 1975; Lewis and Hodgkin 1977), including direct screens for chemotaxis-defective (che and tax) and thermotaxis-defective (ttx) mutants. Animals with defective responses to volatile odorants, but not water-soluble attractants (odr mutants), and animals with defects in osmotic avoidance (osm mutants) have also been identified (Culotti and Russell 1978; Bargmann et al. 1993; J.H. Thomas, pers. comm.). Chemotaxis-defective mutants have also been identified in morphological screens for animals with deformed sensory cilia. Exposed sensory neurons of the amphid and phasmid take up fluorescent dyes including FITC and DiO; animals that do not take up these dyes (dyf mutants) have defects in the sensory neurons or the support cells associated with the amphids (Perkins et al. 1986; Starich et al. 1995). Since a subset of chemosensory neurons regulate the dauer/non-dauer developmental decision, some mutants that were first detected based on defects in dauer formation (daf) also have chemosensory defects and vice versa (Lewis and Hodgkin 1977; Albert et al. 1981; Riddle et al. 1981). The class of mutants that fails to take up FITC includes genes named che, osm, dyf, and daf.

Table 3. Chemotaxis- and thermotaxis-defective mutants.

Table 3

Chemotaxis- and thermotaxis-defective mutants.

A. Genes Required for Cilium Structure and Function

Approximately 25 genes are required for normal formation of the amphid sensilla and normal dye filling of the sensory neurons (Perkins et al. 1986; Starich et al. 1995). The chemosensory and mechanosensory neurons are the only ciliated cells in C. elegans, and thus many genes required specifically for cilium function should be detected in these screens. Most of the dye-filling mutants examined by electron microscopy have been shown to have stunted or malformed sensory cilia (Lewis and Hodgkin 1977; Albert et al. 1981; Perkins et al. 1986). One group of cilium structure genes affects all sensory cilia, including the amphid neurons, phasmid neurons, inner labial neurons, and the mechanosensory neurons associated with the inner and outer labial sensory organs and the cephalic sensory organs. Many of these mutants affect particular ciliary structures, such as the distal or medial cilium elements (Perkins et al. 1986). Although the products of these genes have not been identified, it is likely that they encode components of the cilia and molecules required for cilium assembly.

A few of the cilium structure mutants are more selective in the neurons that they affect. Several genes affect the amphid and phasmid neurons, but not other sensory neurons, suggesting a functional similarity between the amphids in the head and the phasmids in the tail (Perkins et al. 1986). In osm-3 mutants, only the exposed amphid neurons and the phasmid neurons are abnormal. Even within the amphid, the wing neurons (which sense volatile odorants) and the AFD neurons (which sense temperature) are normal in osm-3 mutants. osm-3 encodes a novel kinesin-related protein (Shakir et al. 1993b). Its similarity to microtubule-directed motors suggests a function in assembly or action of the microtubule-rich cilia of the sensory neurons. osm-3 appears to be expressed in the exposed sensory neurons of the amphid and phasmid, as well as the exposed IL2 neurons of the inner labial sensory organs (Tabish et al. 1995). The IL2 neurons, however, are structurally normal in osm-3 mutants.

Assembly of each sensory organ involves interactions between the sensory neurons and two nonneuronal cell types, the socket and the sheath cells (Ward et al. 1975; Ware et al. 1975). The socket cell attaches the sensory organ to the cuticle and, in the case of chemosensory organs, makes the pore in the cuticle. The sheath cell surrounds the dendrites of the sensory neurons and secretes a matrix around the chemosensory cilia. Some cilium structure genes affect the support cells of the sensory organs, rather than the sensory neurons themselves. Mutations in the daf-6 gene lead to degenerations of the amphid and phasmid sheath cells (Albert et al. 1981). Genetic mosaic analysis indicates that daf-6 acts within the sheath cells (Herman 1984). Other genes, including che-14 , have been suggested to act within the sheath or socket cells on the basis of the morphological defects observed in these mutants by electron microscopy (Perkins et al. 1986). This subset of cilium structure genes may be involved in support cell function or sensory organ assembly (for a discussion of organogenesis in the male tail, see Emmons and Sternberg, this volume).

Most of the cilium structure mutants are pleiotropically defective in chemotaxis to volatile and nonvolatile odorants, avoidance of repellents, and mechanosensation in the nose, behaviors that are mediated by ciliated sensory neurons. These mutants display normal thermotaxis behaviors and normal morphology of the AFD sensory neurons (Perkins et al. 1986). Interestingly, different cilium structure mutants have different effects on dauer larva formation. The most severe cilium-defective mutant, daf-19 , is dauer-constitutive; it forms dauer larvae regardless of food or population density. All cilium structure mutants except daf-19 show some rudimentary or malformed cilia, but daf-19 mutants have no recognizable cilium structures (Perkins et al. 1986). Since killing the sensory neurons leads to a dauer-constitutive phenotype (Bargmann and Horvitz 1991b), these results suggest that the sensory neurons in daf-19 mutants are functionally silent. In contrast, all of the other cilium structure mutants are dauer-defective; they never or rarely form dauer larvae. When the sensory neurons were killed in two mutants of this class che-2 and daf-10 , the mutants did form dauer larvae (Bargmann and Horvitz 1991b). Therefore, the sensory neurons in these cilium structure mutants can prevent dauer formation, but they are no longer regulated correctly by external stimuli.

B. unc-86 and lin-32 Affect Chemotaxis and Thermotaxis and Interneuron Cell Lineages

Some mutants are characterized by the loss of many sensory responses without obvious structural defects in the sensory neurons. Animals mutant for the unc-86 gene lack responses mediated by the AWC, ASE, and AFD sensory neurons (chemotaxis and thermotaxis), but they retain at least some function of the AWA sensory neurons (chemotaxis) and the ASH sensory neurons (avoidance) (Bargmann et al. 1993; Mori and Ohshima 1995; C. Bargmann, unpubl.). UNC-86 protein is not expressed in amphid sensory neurons, but it is expressed in the AIZ interneurons, which are directly implicated in thermotaxis and probably important for chemotaxis as well (Finney and Ruvkun 1990; Mori and Ohshima 1995). Lineage changes in unc-86 mutants prevent the generation of the AIZ neurons (Chalfie et al. 1981). unc-86 encodes a POU-class homeodomain-containing transcription factor that controls cell lineage and cell fate during development (Finney et al. 1988; see Ruvkun, this volume).

A second transcription factor required in the lineages that give rise to the AIZ interneurons is encoded by the lin-32 gene, which is a C. elegans homolog of the achaete-scute family of neurogenic genes (Zhao and Emmons 1995). lin-32 mutants have multiple lineage defects and defects in survival and differentiation of neurons, but their amphid chemosensory neurons are present and morphologically normal (Zhao and Emmons 1995; C. Kenyon and E. Hedgecock, pers. comm.; C. Bargmann, unpubl.). lin-32 mutants display broad chemotaxis defects; they are probably defective in processing of chemosensory information by the AIZ neurons and other neurons (Bargmann et al. 1993).

C. tax-2, tax-4, daf-11, and daf-21 Affect Multiple Sensory Behaviors and Chemosensory Axon Outgrowth

tax-2 , tax-4 , daf-11 , and daf-21 mutants have behavioral defects that are similar to those of unc-86 mutants. All of these mutants have defects in AWC-mediated and ASE-mediated chemotaxis, but not AWA-mediated chemotaxis (Dusenbery et al. 1975; Bargmann et al. 1993; Vowels and Thomas 1994; C. Coburn et al., in prep.). In addition, all except daf-11 exhibit essentially athermotactic (non-temperature-responsive) phenotypes in thermotaxis (Dusenbery et al. 1975; Hedgecock and Russell 1975; I. Mori et al., unpubl.).

Animals mutant for these four genes also have characteristic axon outgrowth defects (C. Coburn et al., in prep.), in which some chemosensory axons are elongated or inappropriately branched compared to those of wild-type animals. By inference, the normal function of these genes limits or inhibits axon outgrowth of those neurons. Only a few chemosensory neurons show these defects, typically one or two per animal. The axons of many other neurons are normal in these mutants, suggesting that their defects are relatively specific to the sensory systems.

All four of these genes also affect sensory control of dauer larva formation. daf-11 and daf-21 have a strong dauer-constitutive phenotype and form nearly all dauer larvae at high temperatures (Riddle et al. 1981; Thomas et al. 1993). daf-11 and daf-21 have been proposed to act in the sensory neurons early in dauer formation (Vowels and Thomas 1994); the appearance of a sensory axon phenotype in these mutants supports this model. tax-2 and tax-4 have a slight dauer-constitutive phenotype (C. Coburn et al., in prep.). In addition, tax-2 and tax-4 strongly suppress the dauer-constitutive phenotype of daf-11 and daf-21 , indicating that they can play an important part in dauer larva formation (C. Coburn et al., in prep.).

Temperature-sensitive alleles of tax-2 and daf-11 have been used to define the times at which these genes act to influence axon guidance and chemosensory behavior (C. Coburn et al., in prep.). The chemosensory neurons are born in the embryo and function by the first larval stage. However, the tax-2 and daf-11 gene products are required until the adult stage to maintain normal axon morphology in the adult. Conversely, providing these gene products as late as the fourth larval stage can rescue the axon guidance defects in the mutants. These results indicate that axon morphology is plastic long after the initial neuronal connections are made.

daf-11 encodes a predicted transmembrane guanylyl cyclase, and tax-2 and tax-4 encode two potential subunits of a cyclic-nucleotide-gated channel (D. Birnby and J.H. Thomas, pers. comm.; C. Coburn and C. Bargmann; H. Komatsu et al.; both unpubl.). These sequence similarities suggest that cGMP is an important second messenger in chemosensory and thermosensory transduction and in chemosensory axon morphology. Cyclic-nucleotide-gated channels are implicated in olfactory and photosensory transduction in vertebrates (Fesenko et al. 1985; Nakamura and Gold 1987), and thus this function may be conserved across diverse sensory systems.

D. Genes Required for the Function of Specific Sensory Neurons

Some chemosensory mutants display defects in only a few cell types. For example, the behavioral defects in che-1 , che-6 , che-15 , and che-16 mutants are mostly limited to functions of the ASE sensory neurons (Ward 1973; E. Troemel and C. Bargmann, unpubl.), and the genes osm-7 , osm-8 , osm-11 , and osm-12 appear to affect the functions of the ASH sensory neurons (J.H. Thomas, unpubl.). In these cases, one cell type is functionally silent but morphologically intact, although subtle morphological defects have been observed in che-1 amphid neurons (Lewis and Hodgkin 1977). These genes might affect the specification of particular cell types in development, or they might be required for cell-type-specific neuronal functions.

One gene required for the function of a specific olfactory neuron type is odr-7 (Sengupta et al. 1994). odr-7 null mutants lack all function of the AWA olfactory neurons, but the AWA neurons are present and morphologically normal in the mutants. odr-7 encodes a protein with a zinc finger DNA-binding domain related to those of the steroid receptor superfamily. Unlike other members of this family, it lacks a typical steroid- or ligand-binding domain. Expression of odr-7 is limited to the AWA neurons, apparently within their nuclei, supporting the idea that it encodes a transcription factor involved in AWA function.

What is the role of odr-7 in the AWA neurons? A missense mutation in the odr-7 gene shows an interesting partial loss of AWA function (Sengupta et al. 1994). AWA neurons detect both diacetyl and pyrazine, and the missense mutants are defective only in the response to diacetyl. The clean separation of diacetyl and pyrazine responses argues that odr-7 does not simply specify a cell as AWA or non-AWA. Instead, this result suggests that the odr-7 gene product helps determine the sensory specificity of the AWA olfactory neurons, perhaps by regulating olfactory receptors or other olfactory signaling molecules within the AWA neurons. This model is supported by the observation that expression of odr-10 , the potential diacetyl receptor, is controlled by odr-7 (see Section H below; Sengupta et al. 1996).

E. A Glutamate Receptor Participates in Integration of Sensory Information

An interesting class of mutants lacks some, but not all, of the responses mediated by one type of sensory neuron. The olfactory mutants exemplified by odr-2 and odr-4 have defects in a subset of olfactory responses mediated by one neuron type (Bargmann et al. 1993). For example, strong odr-4 mutants do not respond to diacetyl (sensed by AWA), but they do respond to pyrazine (also sensed by AWA). These selective defects would be expected of genes that are involved in olfactory recognition and signal transduction within the olfactory neurons.

Alternatively, genes in this group might be involved in downstream integration of sensory information, as has been observed for a C. elegans glutamate receptor gene (see Driscoll and Kaplan, this volume). Null mutations in glr-1 , a putative AMPA-type glutamate receptor, lead to a specific defect in avoidance of nose touch (Hart et al. 1995; Maricq et al. 1995). Nose touch is detected by the ASH sensory neurons (Kaplan and Horvitz 1993), which also detect high osmotic strength, but avoidance of high osmotic strength is normal in glr-1 null mutants. glr-1 is expressed in the interneurons that are postsynaptic to the ASH sensory neurons (Hart et al. 1995; Maricq et al. 1995), and mosaic analysis indicates that the gene acts in the interneurons (Hart et al. 1995).

Taken together, these results suggest that nose touch and high osmotic strength generate distinct signals to the interneurons. One model is that activation of the ASH neurons by nose touch leads to glutamate release, which transmits a signal to the interneurons via the glr-1 receptor. Activation of the ASH neurons by high osmotic strength can transmit a signal to the interneurons without using the glr-1 receptor. This might occur if the ASH neuron contains two different neurotransmitters, or if the downstream neurons express multiple classes of transmitter receptors, or if complex signals from multiple sensory neurons contribute to the response.

F. Candidate Receptors for Chemosensation

A family of G-protein-coupled seven-transmembrane domain receptors are thought to be the olfactory receptors in vertebrates (Buck and Axel 1991), although their ligands are unknown. Similarly, a group of more than 40 seven-transmembrane domain receptors include strong candidates for chemosensory receptors in C. elegans (Troemel et al. 1995). These genes, which are called sra, srb, srg, srd, sre, and sro genes, are unrelated in sequence to other known seven-transmembrane receptors, with the exception of sro-1 , which is distantly related to opsin genes. At least 11 of these genes are expressed in subsets of chemosensory neurons, but 3 are expressed outside the chemosensory system. Therefore, although some of these genes might encode chemosensory receptors, others probably have different functions.

As yet, there is no functional evidence that these genes interact with chemoattractants in the sensory neurons. However, their sequences and their expression patterns are consistent with such a role. In addition, their expression patterns fit predictions about chemosensory receptors made by previous analyses of chemosensation in C. elegans. For example, saturation and cross-adaptation patterns in chemosensation indicate that multiple receptors might be expressed in one sensory neuron (Bargmann et al. 1993; Colbert and Bargmann 1995). Consistent with this model, one neuron (ASK) can express at least four different receptors of the sra and srg classes. In addition, laser ablation experiments indicate that a single attractant can be detected by two or more classes of sensory neurons, and some sra, srb, and sre genes are expressed in several different chemosensory neurons. Three different genes, sra-1 , sra-6 , and srd-1 , are expressed in male-specific chemosensory neurons (Troemel et al. 1995). Therefore, these genes are candidate receptors for pheromones during male/hermaphrodite mating in C. elegans. The cells that have been shown to express sr genes detect water-soluble attractants, repellents, or pheromones; no expression of these genes has been observed in the AWA, AWB, or AWC neurons that detect volatile molecules.

G. odr-10 Defines a Potential Olfactory Receptor-Odorant Interaction

odr-10 mutants were isolated in behavioral screens for animals that fail to chemotax to the odorant diacetyl (Sengupta et al. 1996). Interestingly, their chemotaxis to all other tested odorants was found to be normal. odr-10 encodes a potential seven-transmembrane domain receptor that is distinct from the candidate sr receptors, although it has a distant similarity to the srd genes as well as a distant similarity to some vertebrate candidate olfactory receptors. At least ten additional genes that are similar to odr-10 have been identified by the C. elegans genome sequencing consortium (R. Waterston et al., pers. comm.). Reporter gene fusions with odr-10 are expressed only in the AWA neurons, which detect diacetyl, and expression of odr-10 is regulated by odr-7 , the putative AWA transcriptional regulator (Sengupta et al. 1996). A tagged ODR-10 protein appears to be localized to the AWA sensory cilia. These results are consistent with a sensory function for odr-10 .

Its sequence, expression pattern, and mutant phenotype suggest that odr-10 might encode an olfactory receptor for diacetyl. These observations provide a first view of a defined odorant-receptor interaction in vivo. On the basis of genetic analysis, odr-10 function is highly odorant-specific: Even odorants that are very similar in structure to diacetyl appear to be recognized normally in odr-10 mutants. It is possible that other members of the odr-10 gene family encode receptors for other odorants.

H. ttx-1 Affects Thermotaxis and AFD Sensory Ending Morphology

Screens for thermotaxis-defective mutants have yielded some genes with specific functions in the thermosensory pathways. ttx-1 mutants are cryophilic and severely defective in isothermal tracking on a thermal gradient (Hedgecock and Russell 1975; I. Mori and Y. Ohshima, unpubl.), but they show normal or nearly normal responses to a variety of chemical stimuli (Dusenbery and Barr 1980).

The AFD thermosensory neurons have a rudimentary cilium accompanied by a complex microvillar or brush-like structure at their sensory endings. This sensory structure is absent in ttx-1 mutants, suggesting that it is important for thermosensation (Perkins et al. 1986). At this point, the molecular mechanism of thermosensation is unknown. However, the brush-like structure might contribute to thermosensation by increasing membrane surface area if membrane fluidity helps detect temperature or it might change its shape at different temperatures and act as a thermometer. Alternatively, thermosensation could be initiated by thermoreceptor proteins in the AFD projections. A precedent for a receptor-type thermosensor exists in bacterial thermosensation: In Escherichia coli, the four transmembrane chemoreceptors can also function as thermoreceptors via a temperature-dependent conformational change (Nara et al. 1991). ttx-1 mutants are also hypersensitive to dauer pheromone (Golden and Riddle 1984a). Since the dauer pheromone sensitivity of wild-type animals increases at high temperature, the pheromone hypersensitivity of ttx-1 mutants may imply that at least part of the thermosensory system responsible for thermotaxis is also used in dauer formation (Golden and Riddle 1984b).

Mutations in ttx-2 and ttx-3 genes result in a cryophilic phenotype that is similar to that of ttx-1 mutants (I. Mori et al., unpubl.). The molecular characterization of these ttx genes might reveal components of the thermosensory signal transduction pathway.

Copyright © 1997, Cold Spring Harbor Laboratory Press.
Bookshelf ID: NBK19985