To identify genes regulating the development of the six touch receptor neurons, we screened the F2 progeny of mutated animals expressing an integrated mec-2::gfp transgene that is expressed mainly in these touch cells. From 2638 mutated haploid genomes, we obtained 11 mutations representing 11 genes that affected the production, migration, or outgrowth of the touch cells. Eight of these mutations were in known genes, and 2 defined new genes (mig-21 and vab-15). The mig-21 mutation is the first known to affect the asymmetry of the migrations of Q neuroblasts, the cells that give rise to two of the six touch cells. vab-15 is a msh-like homeobox gene that appears to be needed for the proper production of touch cell precursors, since vab-15 animals lacked the four more posterior touch cells. The remaining touch cells (the ALM cells) were present but mispositioned. A similar touch cell phenotype is produced by mutations in lin-32. A more severe phenotype; i.e., animals often lacked ALM cells, was seen in lin-32 vab-15 double mutants, suggesting that these genes acted redundantly in ALM differentiation. In addition to the touch cell abnormalities, vab-15 animals variably exhibit embryonic or larval lethality, cell degenerations, malformation of the posterior body, uncoordinated movement, and defective egg laying.
THE production of fully differentiated neurons requires not only the generation of the appropriate cell type, but also cell migration and process elaboration. All these developmental events are reflected in the maturation of the set of six touch receptor neurons that sense gentle touch along the body in Caenorhabditis elegans (Chalfie and Sulston 1981). Four of these cells (ALML/R, PLML/R) arise before hatching and have anteriorly directed processes; the remaining two (AVM and PVM) are postembryonic and extend processes toward the ventral cord before proceeding anteriorly within it. During embryogenesis, the ALM cells migrate substantial distances posteriorly (Sulstonet al. 1983). In the first larval (L1) stage, cells in the QR and QL lineage, which give rise to the AVM and PVM cells, respectively, undergo long-range migrations in opposite directions (Sulston and Horvitz 1977).
Touch cell fate is specified through the combinatorial action of genes that affect many cell types (Mitaniet al. 1993). Two genes encoding transcription factors, lin-32 and unc-86, are required in the touch cell lineages so that the touch cell precursors are generated (Chalfie and Sulston 1981; Chalfieet al. 1981; Chalfie and Au 1989). lin-32 is a basic helix-loop-helix gene that is most similar to atonal in Drosophila (Zhao and Emmons 1995), and unc-86 is a POU-type homeobox gene (Finneyet al. 1988). unc-86 activates touch cell-specific gene expression by activating mec-3, a LIM homeobox gene (Way and Chalfie 1988; Mitaniet al. 1993; Xueet al. 1993). UNC-86 and MEC-3 bind to the mec-3 promoter cooperatively to activate and to maintain mec-3 expression (Xueet al. 1993; Lichtsteiner and Tjian 1995). UNC-86 and MEC-3 also bind to the promoters of downstream genes needed for the function of the touch cells (Dugganet al. 1998). mec-17 helps maintain mec-3 expression in late larval stages and in adults (Way and Chalfie 1989; Mitaniet al. 1993). lin-14, a gene that regulates general cell lineage progression in L1 and L2 larvae (Ambros and Horvitz 1984; Ruvkun and Giusto 1989), also acts as a positive regulator of touch cell fate (Mitaniet al. 1993). Several genes (lin-4, egl-44, egl-46, sem-4, and pag-3) appear to restrict touch cell fate to the six cells seen in wild-type animals (Desaiet al. 1988; Mitaniet al. 1993; Wightmanet al. 1993; Basson and Horvitz 1996; Jiaet al. 1996). These negatively acting factors, like the positive factors unc-86 and mec-3 (Finney and Ruvkun 1990; Way and Chalfie 1989), are expressed in many types of neurons. lin-14, sem-4, pag-3, egl-44, and egl-46 all encode nucleoproteins (Wightmanet al. 1993; Basson and Horvitz 1996; Jiaet al. 1996; Wuet al. 2001). ced-3 and ced-4, genes involved in programmed cell death (Ellis and Horvitz 1986), also help restrict the number of touch-like cells (Mitaniet al. 1993).
The positions of four of the touch cells (AVM, PVM, and ALMR/L) are determined by several genes that influence the migrations of these cells and their precursors. As with the genes governing cell fate, these genes act in many other cell types and regulate diverse aspects of development (Branda and Stern 1999). Most is known about the genetic control of the Q cell migrations. The QR and QL cells arise at equivalent ventral positions on each side of the animal. The first detectable difference in the behavior of these cells in wild-type animals is that QR migrates dorsally and anteriorly, whereas QL migrates dorsally and posteriorly. Two Antennapedia class homeobox genes, lin-39 and mab-5, influence the positioning of AVM and PVM because they are needed for Q cell migrations (Chalfieet al. 1983; Kenyon 1986; Costaet al. 1988; Salser and Kenyon 1992; Clarket al. 1993; Wanget al. 1993). lin-39 is needed for QR and its descendants to migrate anteriorly through the midbody region. mab-5 is expressed in QL and is needed for QL descendants to migrate posteriorly. Five genes (mig-1, mig-14, egl-20, lin-17, and bar-1) encode components of a Wnt signaling pathway that activates mab-5 expression in the QL neuroblast during its migration (Harriset al. 1996; Maloofet al. 1999; S. Clark and C. Bargmann, personal communication). Seven other genes (egl-27, hch-1, mig-5, unc-11, unc-40, unc-73, and mig-13) also affect the migrations of Q descendants (Hedgecocket al. 1987; Symet al. 1999). Seven genes (egl-27, lin-32, mig-2, mig-10, mig-14, unc-73, and vab-8) significantly influence the embryonic migration of the ALM cells (Hedgecocket al. 1987; Chalfie and Au 1989; Manser and Wood 1990; Harriset al. 1996; Runet al. 1996; Wightmanet al. 1996; Manseret al. 1997; Wolfet al. 1998; Hermanet al. 1999).
No genes have been identified that regulate the outgrowth of only the touch cell processes. Twenty-five genes, however, originally identified by mutations that made animals uncoordinated (Brenner 1974), affect the outgrowth of these processes (Siddiqui 1990). These affect both ventral outgrowth of the AVM and PVM cells (e.g., unc-6 and unc-40) and longitudinal outgrowth from the ALM and PLM cells (e.g., unc-23 and unc-53).
The previous screens for touch-insensitive mutants did not identify mutations affecting touch cell outgrowth or position. Mutations affecting these events in touch cell development have been found during the characterization of mutants isolated for different phenotypes. To search for additional mutations affecting these events, we have marked cells with green fluorescent protein (GFP) to examine the cells in living animals. In this article we describe the identification of several unusual alleles of previously identified genes as well as mutations in two new genes that are needed for touch cell differentiation.
MATERIALS AND METHODS
Strains: Strains were grown as described by Brenner (1974). The following mutations and genetic elements were used in these experiments. They are listed in Appendix 1 of Riddle et al. (1997), except for uIs10[mec-9::lacZ] (Duet al. 1996).
LG I: mig-1(e1787), lin-6(e1466), unc-11(e47), unc-73(e936), dpy-5(e61), unc-40(e271)
LG III: mec-12(u67), mab-21(bx53), dpy-17(e164), unc-86(e1416), dpy-19(e1259), unc-32(e189), tra-1(e1099), dpy-18(e364)
LG IV: mec-3(u6), dpy-20(e1282)
LG V: dpy-11(e224), unc-51(e369)
LG X: dpy-3(e27), lin-32(u282), unc-20(e112), lon-2(e678), mec-2 (e75), mec-2(u299), unc-6(e78), dpy-7(e1324), unc-18(e81), mec-7 (u428), dpy-6(e14), unc-110(e1913e2383), unc-58(u495n1273), unc-115(mn481), sma-5(n678), yDf3, szT1, stDp2, uIs10
Integration of a mec-2::gfp transgene: Fifty L4 larvae or young adults of TU2168, a strain of wild-type background with an extrachromosomal array (uEx217) of the mec-2::gfp plasmid TU#200 and the rol-6(su1006) plasmid pRF4 (Huang et al. 1995), were irradiated for 5.5 min with a 137Cs source (723 rads/min). One line with an integrated element (uIs9) was isolated from 1000 F2 progeny. The uIs9 line was outcrossed four times, and the insertion was mapped within 1 m.u. of unc-23 in the central cluster of chromosome V [none of 42 Unc progeny from unc-23(e25)/uIs9 exhibited GFP fluorescence]. In uIs9 animals, all six touch cells express mec-2::gfp brightly. In addition, a few other cells (mainly in the head) exhibit weaker mec-2::gfp expression (Huanget al. 1995). uIs9 animals are Mec (touch insensitive) at both the head and the tail, but the touch cell positions, morphology, and outgrowth are normal in these animals.
Mutant isolation: uIs9 animals were mutated with ethyl methanesulfonate (EMS) as described by Brenner (1974). F1 progeny from mutagenized parents were transferred onto single plates. Six to 12 F2 animals from each F1 were examined using a 40× objective on a Zeiss (Thornwood, NY) fluorescence microscope. Animals with defects in the number, position, or morphology of the touch receptor neurons were recovered. For further characterization, these mutations were outcrossed and separated from uIs9.
Mapping: Linkage analysis against dpy markers for autosomes, three-factor crosses, and complementation tests were performed according to Brenner (1974). For some mapping we used polymorphic sequence-tagged sites (STS) that identified insertion sites of the transposable element Tc1 in the RW7000 Bergerac strain of C. elegans, following the procedure of Williams (1995). Specifically, homozygous mutants segregated from Bergerac/Bristol heterozygous animals were assayed for the presence of STS markers using single worm PCR. Males with X-linked mutations often do not mate. In such cases, heterozygous animals in a tra-1 background were used for mating to produce trans-heterozygous strains for mapping (tra-1 XX animals develop as males; Hodgkin and Brenner 1977).
For seven mutations preliminary mapping data suggested they were alleles of previously known genes. All seven failed to complement known genes and were not mapped further. These mutations were lin-32(u779), unc-86(u780), mec-3(u778), mig-1(u777), unc-40(u786), unc-73(u782), and unc-51(u783). The first three mutations were tested for complementation on the basis of their phenotypes and chromosomal linkage. The next three mutations, all of which were on chromosome I, were mapped relative to unc-11(e47) dpy-5(e61) [unc-11(e47) 3/63 mig-1(u777) 60/63 dpy-5(e61), unc-11(e47) 0/16 dpy-5(e61) 16/16 unc-40(u786) (only Dpy recombinants used), unc-11(e47) 4/22 unc-73(u782) 18/22 dpy-5(e61)], and then tested for complementation with the indicated genes. The unc mutation (u783) was located on chromosome V to the right of the STS marker stP128; it failed to complement unc-51(e369).
The X-linked dominant mec mutation u784 mapped to the right of unc-6 [all 20 Lon progeny from lon-2(e678) unc-6(e78) +/+ + mec-7(u784) were Mec] and to the left of dpy-6 [all 18 Unc progeny and none of the 24 Dpy progeny from + dpy-6(e14) unc-115(mn481)/mec-7(u784) + + were Mec]. This position suggested that u784 was a mec-7 allele. This identification was confirmed by sequencing the mec-7 gene. DNA from u784 animals lacks 2 bp [GA; either bp 1213 and bp 1214 or bp 1214 and bp 1215 by the genomic numbering in Savage et al. (1989)]. This defect causes a frameshift after Met233.
mig-21(u787) mapped to position -4.2 on chromosome III. The strain + mig-21(u787) +/mec-12(u67) + dpy-17(e164) was constructed and the progeny of the Dpy and Mec recombinants segregated from these animals were examined for PVM migration defects by anti-MEC-7 immunofluorescence (Mitaniet al. 1993). All seven Mec recombinants and four of nine Dpy recombinants produced Mig animals. mig-21(u787) was tested against three STS markers on chromosome III: stP19, stP127, and stP17. mig-21(u787); uIs9(/+) animals (identified by defects in PVM and AVM migrations) were collected from the F2 progeny of a mating of the polymorphism-containing strain RW7000 with mig-21(u787); uIs9. Among 37 mig-21 homozygotes, 14 had only stP17, 2 had stP17 and stP127, and none had stP19. u787 is thus located left of stP127 and very close to stP19 (-5.05 m.u.). mig-21(u787) complemented dpy-19(e1259) and mab-21(bx53).
vab-15(u781) mapped to position 1.56 on the X chromosome. None of the 13 Unc recombinants and 12 of the 14 Dpy recombinants from + vab-15(u781) +/dpy-6(e14) + unc-115(mn481) animals segregated u781 animals. The recessive vab-15(u781) mutation complemented unc-58(u495n1273) and unc-115(e1913e2383) and was covered by stDp2.
lin(u788) was difficult to map but appeared to be on chromosome I, since it showed linkage to dpy-5(e61) but not to dpy mutations on other chromosomes.
Germline transformation and rescue: Cosmid and plasmid DNA were purified using the QIAGEN (Chatsworth, CA) Midiprep kit. Germline transformations were performed according to Mello et al. (1992). The rol-6(su1006) plasmid pRF4 (Melloet al. 1992) was coinjected as a dominant marker for transformation. Rescue was scored among rollers. Plasmids were injected at 10-15 ng/μl. Cosmid R07B1 injected at 60 ng/μl fully rescued vab-15(u781) animals. A 15-kb SalI-SpeI fragment from R07B1 was cloned into the SalI and SpeI site of pBSKII (+) to generate plasmid TU#516. The subsequent subcloning all employed the vector pBSKII (+). TU#517 contains the 10-kb XhoI-HpaI fragment of TU#516. TU#518 contains the HindIII-SpeI fragment from plasmid TU#517. TU#519 contains the XbaI fragment from TU#517. TU#520 contains the HindIII-SacII fragment from TU#517. TU#517 also fully rescued u781. A total of 70-80% of the animals in four stable lines were rescued completely, judged by movement, growth, and touch sensitivity of the animals. TU#518 and TU#520 rescued partially since touch sensitivity in the head was often not restored. TU#519 rescued u781 even more poorly.
Sequencing vab-15(u781): Coding regions of genomic DNA for vab-15 were amplified with primers PR-F80 (5′-CGCATT CAGTGTTCCCTCATCCTTTGCAG) and PR-F81 (5′-CGAAAGGCAACCTCTCTAGTGGAGTCAC). All primers used in this work were synthesized by Operon Technology (Alameda, CA). KlenTaq polymerase mix (CLONTECH, Palo Alto, CA) was used for single worm PCR reactions performed as described (Williamset al. 1992). Independent PCR reactions for nine single animals were combined and direct-PCR sequencing was performed with the Sequenase PCR product sequencing kit (United States Biochemical, Cleveland). Additional primers used for sequencing were as follows: PR-F79 (5′-CCTGCTGATTGCTTCTGCTA), PR-F82 (CAATGGTGTC GAAAGATGAAAAACCAAAGC), PR-F83 (5′-GGTGGGTATTTCTGTCACCGCGGAAAAG), PR-F84 (5′-CTATGACTCCA TACTTCGGCGC), PR-F85 (5′-CTTCAATGGCTATCATCTCA GCG), PR-F86 (5′-GTTGCGGAATTTGATTTCACAGC), PR-F87 (5′-CCATAATTATGAGGAGCGTGTCAAG), and PR-F88 (5′-GAGTAGAGAATGGGGTACGTGGC).
Random-primed cDNAs were synthesized from wild-type poly(A)+ RNA isolated as described (Duet al. 1996). We used KlenTaq polymerase mix with primers PR-F82 and PR-F83 to generate vab-15 cDNA by reverse transcriptase (RT)-PCR. cDNA fragments were sequenced with the Sequenase PCR product sequencing kit (United States Biochemical). The primers used for sequencing were PR-F82 to PR-F85 and PR-F88.
Reporter fusions: The SalI-PstI fragment of TU#517 was cloned into pBSKII (+) to yield TU#521. The SalI-SmaI fragment of U#521 was cloned into pPD95.750 (a GFP expression vector designed for C. elegans provided by Andrew Fire) between SalI and SmaI sites to give TU#523. uEx351 and uEx352 were generated by injecting TU#523 and pRF4 into wild-type animals or u781 animals, respectively. TU#523 fully rescued the defects of u781 animals: 70-80% of the animals in the stable line were fully touch sensitive and had wild-type movement.
Mutant isolation: Using a strain with an integrated mec-2::gfp transgenic array (uIs9) to visualize touch cell bodies and processes, we screened F2 progeny representing 2638 EMS-mutagenized haploid genomes for mutants with touch cell defects. We identified 11 mutations in 11 genes (Figure 1). Five mutations affected touch cell fate, 3 affected touch cell migrations, and 3 affected touch cell process outgrowth. Two mutations defined new genes, vab-15 and mig-21.
Mutations defective in touch cell differentiation: Three of the five mutations that affected touch cell fate failed to complement mutations in previously identified genes: lin-32(u779), unc-86(u780), and mec-3(u778) (Figure 1B). lin-32(u779) animals lacked the AVM and PVM cells. In addition, the ALM cells failed to migrate posteriorly and had cell bodies that were close to the pharynx. These phenotypes have been seen in lin-32 animals with other partial loss-of-function mutations (Chalfie and Au 1989; Zhao and Emmons 1995). One-third of the u779 animals (without uIs9) were touch insensitive (Mec) at the head; an additional one-sixth of the animals were partially Mec at the head. Portman and Emmons (2000) recently found that u779 animals are hypomorphic with regard to the development of the male rays despite finding that the mutation was a very early nonsense mutation. unc-86(u780) animals were Mec and had only two cells expressing mec-2::gfp. These cells were located in the tail and their processes were longer than those of wild-type PLM cells. Similarly, only two cells expressed lacZ from the touch cell-specific mec-9 promoter in the integrated element uIs10 in u780 animals. These results suggest that these two cells exhibit a partial touch cell fate and may be nonfunctional PLM cells. In contrast, unc-86(e1416) animals do not detectably express mec-7 (Mitaniet al. 1993) or mec-9::lacZ in any cells. The u780 mutation is probably a partial loss-of-function allele. Mutants with a temperature-sensitive (partial loss-of-function) unc-86 mutation also had tail cells that expressed a touch cell marker (mec-7::lacZ) at the restrictive temperature (Hamelinet al. 1992).
mec-3(u778) animals were Mec. Only the PLM cells expressed mec-2::gfp in these animals and this expression was weak. Similar weak expression of genes required for touch cell functioning has been seen in other mec-3 strains (Mitaniet al. 1993), including a strain with the mec-3(u6) mutation, a nonsense mutation that causes a truncation of the protein in the middle of the homeodomain (Xueet al. 1993).
The remaining two mutations affecting touch cell fate were u781 and u788. u781 is described below. lin(u788) I is a temperature-sensitive mutation that proved difficult to map because of its near sterility at 25°. Twenty-three of 37 u788 animals at 25° had at least one additional AVM-like or PVM-like cell. The defect was seen in only 1 of 32 animals raised at 15°. At the higher temperature the animals are also egg-laying defective.
vab-15(u781) has pleiotropic defects: u781 animals lacked the AVM, PVM, and PLM cells (Table 1). In addition, at least one of the ALM cells often failed to migrate or migrated a shorter distance and remained close to the pharynx. u781 is a mutation in a previously uncharacterized gene, which we have named vab-15 because of the variable morphological defects seen in u781 mutants. In addition to the absence of four touch cells, vab-15(u781) animals exhibited severe developmental defects. u781 was partially lethal; approximately two-thirds of the animals failed to survive. Most of the animals died as embryos around the twofold stage; others died in different larval stages. Before the animals died, extensive cell deaths occurred in all cell layers but especially in epidermal cells in both embryonic and larval animals (Figure 2, c and d). Dying cells formed large vacuoles that accumulated and persisted in the dying animals. The morphology of the dying cells was distinct from that of programmed cell death (Sulston and Horvitz 1977), but similar to the vacuolated deaths produced by gain-of-function mutations in the degenerin genes mec-4 and deg-1 (Chalfie and Sulston 1981; Chalfie and Wolinsky 1990). The vab-15 deaths, however, differed from those produced by the degenerin mutations, since they were not suppressed by mec-6 mutations (Chalfie and Wolinsky 1990). Cell death occurred less frequently in the head than in the body. The late embryos and L1 animals had severely deformed posterior bodies that were often shortened dramatically (Figure 2, b and c). Adult hermaphrodites had variably enlarged and shortened tails, and the body cuticle was twisted as in rollers. u781 animals also had a severe egglaying defect, and some had protruding vulvae. The animals did not respond to the touch of an eyebrow hair or to the harsher prod of a platinum wire (the Tab phenotype), and they were uncoordinated (Unc; they coiled and kinked when attempting to move backward).
vab-15 is a msh-type homeobox gene: We mapped vab-15 to the X chromosome at position 1.56. We found a 9.5-kb SalI-HpaI fragment (TU#517) from cosmid R07B1 that fully rescued all the morphological and behavioral defects in vab-15(u781) animals (Figure 3). The region was sequenced by the C. elegans Genome Project (C. elegans Sequencing Consortium 1998) who predicted that it contained one gene with four exons (GenBank accession no. Z48621). We isolated a cDNA (GenBank accession no. AF286218) for vab-15 by RT-PCR and confirmed the predicted splicing pattern of this gene (Figures 3 and 4a). Subclones of the SalI-HpaI fragment (TU#518-TU#520) rescued vab-15 incompletely (Figure 3). Injection of TU#518 or TU#520 resulted in animals in which touch sensitivity at the head was often not restored. Transformation with TU#519 rarely restored touch sensitivity, and movement was rescued in very few animals. These results suggest that virtually all of the 9.5-kb SalI-HpaI fragment is required for normal vab-15 expression.
vab-15 is similar to the msh (muscle segment homeobox) class of homeobox genes. msh genes are found in a wide variety of animals, including hydra, Drosophila, sea urchin, ascidian, zebrafish, chicken, Xenopus, mouse, and humans (Davidson 1995; Holland 1991). In vertebrates, msh-like genes were initially named Hox-7 or Hox-8, but are now called Msx (Davidson 1995). The homeodomains of all the msh genes and vab-15 are at least 82% identical; additional regions of similarity extend seven amino acids N-terminal and nine amino acids C-terminal of the homeodomain (Figure 4b). vab-15 was the only msh-like gene found in the entire C. elegans genome (C. elegans Sequencing Consortium 1998). Although multiple Msx genes have been found in vertebrates, only single genes have been reported in invertebrates (Davidson 1995).
We confirmed that this predicted gene was vab-15 by sequencing the u781 allele. The splice donor site in the second intron was ATAAGC instead of GTAAGC. In C. elegans, the “GT” sequence in the splice donor site is invariant (Emmons 1988). This intron precedes the exon encoding the homeodomain of VAB-15.
vab-15 is expressed mainly in embryos and young larvae: To identify cells expressing vab-15, we transformed animals with a plasmid (TU#523) that encodes VAB-15 with GFP fused at its C terminus (Figure 3). TU#523 fully rescued the defects of u781 animals. The rescued animals (uEx352) showed strong expression of gfp in embryos starting from the midgastrula stage (Figure 5). About a dozen nuclei expressed vab-15::gfp in midgastrula (Figure 5a). Before and at the comma stage, numerous nuclei in all cell layers expressed vab-15 (Figure 5, b and c). Embryos approaching the twofold stage had strong GFP fluorescence in ectodermal cells, including hypodermal cells and neuroblasts (data not shown). vab-15 was expressed strongly in the set of 12 P cells from before hatching and during the L1 stage (Figure 5d). In L2 and early L3 animals vab-15 was also expressed in seam cells and ventral cord motor neurons (Figure 5, e and f). Five unidentified cells in the head and four cells in the tail (PHBL/R and PVCL/R) expressed vab-15 throughout larval development. These cells were the only ones expressing the fusion in L4 larvae and adults (data not shown). vab-15 expression was nuclear at all stages. The head was almost devoid of vab-15 expression, a result that is consistent with the finding that most u781-induced defects occurred behind the pharynx. The strong expression in embryos and the embryonic lethality in u781 animals suggest that vab-15 plays an essential role in embryonic development. No fluorescence was seen in the touch cells or any cells of the Q lineages (the lineages that give rise to the AVM and PVM cells).
vab-15(u781) mutants lack posterior mec-3- and unc-86-expressing cells: To investigate the role of vab-15 in the generation of the touch cells, we examined mec-3 and unc-86 expression in vab-15(u781) animals. We used an extrachromosomal array (uEx326) containing a mec-3::gfp fusion gene that was expressed in all six touch cells and the two FLP neurons (S. Luo and M. Chalfie, unpublished data). In u781; uEx326 animals, only the ALM and FLP cells expressed gfp at wild-type levels (Table 1). We examined unc-86 expression in u781 animals with a rabbit anti-UNC-86 antibody (Finney and Ruvkun 1990). UNC-86 was easily detected in the ALM cells and their sister cells, the BDU neurons, but was largely missing from AVM, PVM, or PLM cells (Table 1). These results suggest that vab-15(u781) interfered with the production or differentiation of these unc-86- and mec-3-expressing cells in the posterior of the animal.
In wild-type animals, unc-86 is expressed in 57 cells: 37 cells in the head, 10 in the tail, and 10 in the body (Finney and Ruvkun 1990). In u781 animals, only 3-6 of the 10 tail cells and 6 of the 10 body cells (ALMR/L, HSNR/L, and BDUR/L) that normally express unc-86 did so. In contrast, all 37 unc-86-expressing cells in the head were seen (data not shown). These results further suggest that vab-15 affects development of posterior cells.
vab-15 and lin-32 redundantly activate ALM cell fate: A lin-32 mutation, u282, produces touch cell defects similar to those produced by vab-15(u781): the AVM, PVM, and PLM cells are missing, and the ALM cells are more anteriorly displaced (Chalfie and Au 1989). To see whether these two genes act together in regulating ALM cell fate, we examined touch cells in a lin-32(u282) vab-15(u781) double mutant by mec-7 immunocytochemistry (Table 1). The single mutations cause the displacement, but not the loss, of the ALM cells as seen for uIs9 expression (data not shown) and mec-3 expression (Table 1) with u781 and for mec-7 in situ hybridization (Mitaniet al. 1993). In contrast, 49% of the double mutant animals lacked ALM cells and another 38% had only one ALM cell (N = 45). In addition, when the ALM cells were present, their cell bodies were often anterior to the rear bulb of the pharynx, indicating more severe defects in ALM migration. These additive effects suggest that vab-15 and lin-32 have somewhat redundant roles in activating touch cell fate in ALM cells. Since lin-32 is needed for the generation of the posterior touch cells, by extension, vab-15 may also be needed for their production.
Mutations affecting touch cell migration: We identified three mutations that affected AVM or PVM migrations: mig-1(u777), unc-40(u786), and u787, an allele of a new gene we have named mig-21. Forty-three of 46 mig-1(u777) animals had defects in the position of the PVM cells: in 32 animals PVM was in the same anterior position as AVM, in 7 animals PVM was in the midbody region, and in 4 animals PVM was in the tail just anterior to the anus. For another allele (e1787) of mig-1, we found 63 of 87 animals with PVM migration defects: 51 PVM cells were in the anterior region and 12 were in the midbody region. Similar data were obtained by Harris et al. (1996) with this latter allele.
unc-40 encodes a netrin receptor of the deleted in colon cancer (DCC) family (Hedgecocket al. 1990; Chanet al. 1996). unc-40(u786) animals were shorter than wild type, mildly egg-laying defective (Egl), and uncoordinated (Unc). The AVM cells failed to migrate in 7 of 56 (12%) animals; in 19 animals (34%) PVM migrated anteriorly instead of posteriorly. In 21 of 56 (38%) u786 animals AVM or PVM processes failed to grow ventrally, but instead fasciculated with ALM or PLM axons or simply grew laterally. This outgrowth defect was independent of the migration defect. Hedgecock et al. (1990) found similar defects produced by two other unc-40 alleles, e271 and rh66: AVM migration defects were seen in 2-3% of the animals, PVM migration defects in 20-25%, and AVM outgrowth defects in 19%. u786 seemed to be slightly more severe with regard to these touch cell defects.
mig-21(u787) animals were defective in the placement of both AVM and PVM. PVM was located anteriorly, usually at a position equivalent to that of the wild-type AVM cell, in 51 of 60 animals. AVM was located posteriorly, usually at the same relative position as the wild-type PVM cell, in 6 of 60 animals. These positioning defects originate from defects in the migration of the QL and QR cells (Figure 6a). At hatching, QL and QR are at the same position along the anterior-posterior axis on each side of the animal. In wild-type animals QL migrates posteriorly to above V5 before the first cell division, and its descendants continue to migrate toward the posterior; QR migrates until it reaches P7/8 and divides, and its descendants continue to migrate anteriorly (Figure 6, b and c). We followed the migration of six QL cells and four QR cells in 10 u787 animals. QL and QR behaved rather similarly. They migrated randomly anteriorly or posteriorly, and sometimes they failed to migrate completely. The descendants of all six QL and three of the four QR cells migrated anteriorly. The descendants of the one QR that migrated posteriorly to above V5R also migrated posteriorly (Figure 6, b and c). Thus, mig-21 is required for the initial asymmetric migrations of Q cells, but the subsequent migrations of their descendants appear to depend on the final position of the Q cells. The predominance of anteriorly positioned AVM/PVM cells appears to reflect the finding that few of the Q cells migrate to positions above V5.
Mutations affecting outgrowth of touch cell processes: We found three mutations that affected touch cell outgrowth (Figure 7). All three mutations were in previously identified genes: unc-51 [which encodes a serine/threonine kinase (Oguraet al. 1994)], unc-73 [which encodes a guanine nucleotide exchange factor regulating cell polarity (Runet al. 1996)], and mec-7 [which encodes a β-tubulin (Savage et al. 1989, 1994)]. unc-51(u783) and unc-73(u782) mutants were paralyzed and exhibited similar touch cell defects. Both genes affect axonogenesis and influence the outgrowth of many types of neurons (Mcintireet al. 1992). In either mutant, the ALM processes ended at or before the nerve ring, not near the tip of the nose as they do in wild-type animals. The PLM processes often failed to grow more anteriorly than the eventual position of the vulva. In addition, L1 and L2 larvae often had touch cell processes with enlarged growth cones. In unc-51(u783) animals, the PLM processes often bifurcated, whereas those in unc-73(u782) animals usually turned to join the ventral cord at their ends. The AVM and PVM processes were severely shortened in u783 animals. unc-73(u782) and unc-51(u783) mutations also appear to affect touch cell migration. The AVM cell failed to complete its anterior migration in 9 of 12 u782 animals. In contrast, ALM and AVM cell bodies were much too anterior in all 38 u783 animals examined. This migration defect, albeit weaker, was also seen in unc-51(e369) animals, a phenotype noted by Hedgecock et al. (1985). While ALM cells in 15 otherwise wild-type uIs9 animals were properly positioned, the ALM cells in all 9 unc-51(e369) animals were slightly more anterior. In 3 of these animals one ALM cell was severely mispositioned near the rear bulb of the pharynx. Recently, Branda and Stern (2000) found that while unc-51 mutations do not affect the migration of sex myoblasts in otherwise wild-type animals, they affected these migrations in egl-17 animals. These researchers also found that the e369 allele was hypomorphic with regard to this phenotype.
All touch cell processes in animals with the dominant mec-7(u784) mutation were dramatically shortened and had dense branches at their ends. In young larvae, both PLM and ALM cells have enlarged growth cones with branches. The mutation in u784 is a deletion of two nucleotides [G1213 and A1214 in the genomic sequence of Savage et al. (1994)]. The deletion is at amino acid 233 and the frameshift results in a stop codon 40 amino acids later. The truncated β-tubulin might interfere with the assembly of microtubules (Mitchison and Kirschner 1984; Savageet al. 1994).
We have identified 11 mutations in 11 genes that affect touch cell development. Since each gene is represented by a single allele, we have not saturated the screen for genes that regulate touch cell differentiation, migration, and outgrowth. In addition to identifying two new genes, vab-15 and mig-21, our screen allowed us to easily identify novel or partial phenotypes produced by mutations in several known genes. The lin-32(u779) allele produces partial touch insensitivity at the head and full touch sensitivity in the tail. The lin-32 alleles identified previously produced animals that were either touch sensitive or touch insensitive only in the tail (Chalfie and Au 1989; Zhao and Emmons 1995). The unc-86(u780) mutants have two PLM-like cells in the tail, although these cells express mec-2::gfp at a lower level than do wild-type cells. In contrast, no touch cells are found in e1416, a representative allele of unc-86 (Mitaniet al. 1993). Finally, the mec-7(u784) mutation identified an outgrowth defect resulting from a mutation in a β-tubulin gene.
vab-15 is needed to generate touch cells: Our screen produced mutations in two previously unidentified genes. The vab-15 gene appears to be needed for the generation of the touch cells and several other cells. In vab-15(u781) animals, only the ALM cells expressed unc-86. Since unc-86 is needed for the generation of touch cell precursors (Chalfie and Sulston 1981), the missing touch cells (AVM, PVM, and PLM cells) are probably not generated. vab-15 may affect an earlier stage of touch cell lineages than unc-86 as lin-32 does (Chalfie and Au 1989). We have found that lin-32 and vab-15 have redundant functions in regulating ALM cell fate; animals mutated in both genes had fewer ALM cells and more severe defects in ALM migrations than animals mutated in only one of the genes. Since lin-32(u282) is not a null allele (Zhao and Emmons 1995) and vab-15(u781) may not be, we cannot tell whether the two genes act together or in parallel. The few remaining ALM cells in the double mutant could result from incomplete penetrance of either mutation or from the presence of additional redundant factors that activate ALM cell fate. Regardless, the additive effect of the two mutations suggests that they may act similarly in the development of the precursors for the four touch cells that are missing in both mutants. Since we have not detected vab-15 expression in the Q cells or the touch cells, the gene may affect embryonic cells that give rise to these cells or be necessary for cell interactions that are required for the production of the cells (the expression pattern of lin-32 is not known).
vab-15 plays an essential role in morphogenesis, especially in embryos: The extensive vacuolated cell death in u781 animals probably causes the embryonic and larval lethality. The mechanism of the cell death in vab-15 animals is unknown, but might be due to a general failure to develop a normal body plan, or failure of several cell types to differentiate normally. Vacuolated cell deaths also occur in the embryonic epidermal cell layer in lin-26 mutants in C. elegans (Labouesseet al. 1994). lin-26 is a transcription factor required for epidermal differentiation. In Drosophila, mutation in the polyhomeotic gene causes massive vacuolated cell death in the ventral hypodermis (Smouseet al. 1988). Mutations in the vestigial locus cause similar-appearing cell death in a specific region in the larval wing disc. This cell-autonomous death is due to the failure of these wing disc cells to differentiate (Bownes and Roberts 1981).
The dramatic deformity and lethality suggest that vab-15 plays an essential role in embryonic morphogenesis. The various phenotypes of the surviving animals suggest that the generation or differentiation of several cell types is affected. The Tab phenotype points to defects in interneurons; e.g., animals lacking the PVC neurons are Tab in the tail (Chalfie and Wolinsky 1990). The tail Tab phenotype in vab-15 animals is consistent with the continued expression of vab-15 in wild-type PVC neurons. The Unc phenotype might be caused by defects in motor neurons. The Egl phenotype could be caused by defects in several cells, e.g., the egg-laying muscles, the vulva, or the HSN neurons (Desaiet al. 1988). The HSN neurons in u781 animals, however, express unc-86 and are positioned normally. The deformity in embryos and in the posterior body of adult hermaphrodites may be caused by the inappropriate differentiation or movement of hypodermal cells during embryonic development.
mig-21 regulates the asymmetric migrations of Q neuroblasts: The second new gene uncovered by our screen was mig-21. In mig-21(u787) animals, the asymmetry of the initial Q cell migrations is abolished: the Q cells and their descendants on either side migrate in either direction. This defect appears to be one of establishing the initial asymmetry of the right and left Q cells, since if either cell migrates sufficiently posteriorly (as does the normal QL cell), its progeny will migrate posteriorly. The mig-21 mutation is the only one, so far identified, that affects the initial asymmetry of the Q cells. Mutations in three genes affect the migrations of the Q cells. The QR migration is shortened in unc-73 mutants (Runet al. 1996) and lin-39 mutants (Clarket al. 1993; Wanget al. 1993), and QL occasionally (<10%) migrates anteriorly in egl-5 mutants (Chisholm 1991). Since none of these mutations affect the migrations of both cells, the affected genes probably act downstream of mig-21. The finding that few of the AVM and PVM cells in mig-21 mutants were positioned posteriorly is consistent with the initial migration pattern of the cells in the mutants and the finding that mab-5 is expressed when QL is positioned over the V5L cell (Salser and Kenyon 1992).
Mutations affecting touch cell outgrowth: Three of our mutations caused a dramatic shortening of touch cell processes without affecting the direction of outgrowth. Such defects in adults were reflected in the abnormal enlargement and branching of the growth cones in young larvae. This early phenotype indicates that the abnormal axonal morphology is due to defects in axonal outgrowth. In addition to some general similarities, each of the three mutants has distinct outgrowth defects shown clearly in the PLM cells. PLM processes in unc-73(u782) mutants turned and fasciculated with processes in the ventral cord, PLM processes in unc-51(u783) animals bifurcated, and PLM processes in mec-7(u784) animals branched extensively at their ends. These differences suggest that these genes are needed for distinct aspects of axonal outgrowth.
Both unc-51 and unc-73 mutations affected axonal extension and fasciculation of many types of neurons; they also affected cell migration (Hedgecocket al. 1985; Siddiqui 1990; Siddiqui and Culotti 1991; Mcintireet al. 1992). One intriguing phenotype of unc-51(u783) animals was the severe anterior displacement of ALM and AVM cells. The anterior displacement of AVM in unc-51(u783) animals beyond its normal position suggests that unc-51 does more than simply promote cell migration.
The dominant mec-7(u784) mutation probably affects outgrowth by interfering with microtubule assembly. The ectopic branching caused by this mutation is reminiscent of the production of preterminal (but not more proximal) growth cones when neurons in culture are treated with antimicrotubule drugs (Brayet al. 1978). The truncated product that results from the mec-7(u784) mutation lacks two of three potential interaction domains (Savageet al. 1994) as well as the MAP-binding site at the C terminus. The dominant phenotype, however, suggests that the remaining interaction site allows the mutated MEC-7 to block microtubule assembly.
Screens for touch-insensitive mutants have identified several genes that are needed for the production and function of the touch receptor neurons (Chalfie and Sulston 1981; Chalfie and Au 1989). Since these screens did not produce mutations that affected the outgrowth or migration of the touch cells, these processes were hypothesized to be more general. Our discovery of several genes that affect these events supports this conjecture and suggests that additional screens looking directly at touch cell defects will yield still more genes needed for these aspects of cell differentiation. Forrester et al. (1997), who used a similar approach to look for mutations affecting CAN migrations, were also able to identify a wealth of new genes affecting the CAN cells and many other cells.
We thank Gary Ruvkun for providing the anti-UNC-86 antibody and our labmates for comments on the manuscript. This research was supported by National Institutes of Health grant GM30997 to M.C.
Communicating editor: P. Anderson
- Received September 1, 2000.
- Accepted January 29, 2001.
- Copyright © 2001 by the Genetics Society of America