Co-factor homeodomain proteins such as Drosophila Homothorax (Hth)and Extradenticle (Exd) and their respective vertebrate homologs, the Meis/Prep and Pbx proteins, can increase the DNA-binding specificity of Hox protein transcription factors and appear to be required for many of their developmental functions. We show that the unc-62 gene encodes theC. elegans ortholog of Hth, and that maternal-effect unc-62mutations can cause severe posterior disorganization during embryogenesis (Nob phenotype), superficially similar to that seen in embryos lacking function of either the two posterior-group Hox genes nob-1 and php-3 or the caudal homolog pal-1. Other zygotically actingunc-62 alleles cause earlier embryonic arrest or incompletely penetrant larval lethality with variable morphogenetic defects among the survivors, suggesting that unc-62 functions are required at several stages of development. The differential accumulation of four unc-62transcripts is consistent with multiple functions. The C. elegans exdhomologs ceh-20 and ceh-40 interact genetically withunc-62 and may have overlapping roles in embryogenesis: neither CEH-20 nor CEH-40 appears to be required when the other is present, but loss of both functions causes incompletely penetrant embryonic lethality in the presence of unc-62(+) and complete embryonic lethality in the presence of an unc-62 hypomorphic allele.

INTRODUCTION

Central to the patterning process in all animal embryos are the homeodomain transcription factors known as homeotic selector (Hox) proteins (for a review,see Manak and Scott, 1994). Hox proteins often bind target DNA sites in vitro with little selectivity(Ekker et al., 1994;Hoey and Levine, 1988;Mann, 1995). Their biological specificity appears to depend on interactions with protein co-factors, some of which are members of the TALE (three amino acid loop extension) class of atypical homeodomain proteins. These proteins are further divided into families based on similarities in a more N-terminal domain(Burglin, 1997), which is believed to be important for protein interactions(Abu-Shaar et al., 1999).Drosophila Extradenticle (Exd) and its vertebrate homologs the Pbx proteins, which are characterized by bipartite PBC-A and PBC-B interaction domains, comprise the PBC family. Drosophila Homothorax (Hth) and its vertebrate homologs the Meis/Prep proteins include the Hth/Meis (HM)interaction domain and are referred to as the Meis family (reviewed byMann and Affolter, 1998;Mann and Chan, 1996).

In the PBC family, Exd can associate with and increase the binding specificity of the medial-group Hox proteins Ultrabithorax and AbdA, but not the posterior-group protein AbdB (Chan et al., 1994; van Dijk and Murre,1994). Likewise, the homologous vertebrate Pbx proteins can increase the binding specificity of the first 10 Hox paralog proteins,including the posterior-group Hox9 and Hox10 proteins, but not that of the posterior-group Hox11-Hox13 proteins (Shen et al., 1997b). In the Meis family, the vertebrate Meis/Prep proteins associate with and increase the binding specificity in vitro of Hox9-Hox13 proteins (Shen et al.,1997a).

The PBC- and Meis-family Hox co-factors in turn appear to be co-regulated through subcellular localization (Pai et al., 1998). Exd, which includes both nuclear localization and nuclear export signals in its sequence(Rieckhof et al., 1997), is retained in the cytoplasm after synthesis. Complexing with Hth, however,inactivates the export signal, and the complex becomes localized to the nucleus where it can interact with Hox proteins(Abu-Shaar et al., 1999;Ryoo et al., 1999;Affolter et al., 1999). Analogous studies in vertebrates have demonstrated similar interactions between the Pbx and Meis/Prep proteins(Berthelsen et al., 1999;Ryoo et al., 1999), suggesting that this may be a well-conserved mechanism for regulating Hox gene function.

C. elegans has six Hox genes, three of which have known patterning roles during larval development (Chisholm,1991; Clark et al.,1993; Wang et al.,1993). However, relatively little is so far known about earlier Hox gene functions during embryonic development. In contrast to the situation in Drosophila and vertebrates, only anterior- and posterior-group Hox paralogs, ceh-13 (Brunschwig et al., 1999) and nob-1/php-3(Van Auken et al., 2000),respectively, are essential for embryonic patterning and viability. Neither the medial-group homologs mab-5 and lin-39, nor the posterior-group homolog egl-5 are required for embryonic survival(Chisholm, 1991;Clark et al., 1993;Wang et al., 1993); even embryos lacking function of all three of these genes develop into larvae that exhibit patterning defects but nevertheless survive to become fertile adults(Wrischnik and Kenyon,1997).

In screens for C. elegans embryonic lethal patterning mutants(Edgar et al., 2001), we found a loss-of-function maternal-effect mutation in the unc-62 gene that causes severe posterior defects (Nob phenotype) superficially similar to those of nob-1/php-3 embryos (Van Auken et al., 2000) and embryos lacking zygotic function of thecaudal homolog pal-1(Edgar et al., 2001). We show here that unc-62 is the C. elegans ortholog ofDrosophila hth (previously known as ceh-25)(Burglin, 1997), and that its transcripts include four splice variants predicted to encode proteins with two alternative homeodomains and strong similarity to the Meis family of Hox co-factors. We have molecularly characterized six unc-62 alleles and have analyzed the resulting phenotypes, which range from the originally reported uncoordinated (Unc) phenotype(Brenner, 1974) through larval defects and lethality to early embryonic arrest. Our results suggest multiple functions for unc-62 gene products.

We have also shown that both of the two C. elegans PBC-family members, ceh-20 and ceh-40, interact genetically withunc-62. Although embryos lacking either ceh-20 orceh-40 alone are viable, loss of both functions causes embryonic lethality that is strongly enhanced by a hypomorphic unc-62 allele,suggesting overlapping roles for these three genes during embryogenesis.

MATERIALS AND METHODS

Strains and alleles

C. elegans strains, all derived from the Bristol N2 wild type,were obtained from the Caenorhabditis Genetics Center or from our collection unless otherwise indicated. Stock maintenance, genetic mapping and complementation tests were carried out using standard methods(Sulston and Hodgkin,1988).

Genes and alleles referred to are as follows, by linkage group (LG).

  • LGI: smg-1(r861), glp-4(bn2ts);

  • LGIII: fem-2(b245ts), ceh-20(ay38), nDf16 (a small deficiency that uncovers ceh-20);

  • LGIV: fem-1(hc17ts), him-8(e1489);

  • LGV: dpy-11(e224), unc-46(e177), unc-62(ct344, e644, e917, ku234, s472,t2012), unc-60(m35), and the deficiencies sDf26, sDf27,and sDf50;

  • LGX: ceh-40 (predicted gene; no reported alleles).

The eT1 (III;V) reciprocal translocation was used as a balancer chromosome for unc-62 alleles. The duplication svDp1,derived from sDp3 by fusion with a sur-5::GFP reporter construct (S. Tuck, personal communication), was used to balanceceh-20 alleles.

The canonical unc-62(e644) was originally described by Brenner(Brenner, 1974). Previously undescribed alleles, characterized by failure to complement e644,were obtained as follows: ct344 from our mutant screens(Edgar et al., 2001),ku234 in a screen for defects in vulval morphogenesis (M. S. and M. Han, unpublished), and e917, s472(Rosenbluth et al., 1988) andt2012 as gifts from A. Chisolm, D. Baillie, and R. Schnabel,respectively.

All unc-62 mutations were outcrossed at least three times;s472 was outcrossed ten times prior to phenotypic analysis. Homozygous lethal mutations were balanced over eT1 (III;V) or over LGV doubly marked with unc-60 and dpy-11. The s472deletion mutation was maintained as unc-62(s472) unc-46(e177)/unc-60(m35)dpy-11(e224); presence of the s472 deletion was verified by PCR.

To determine whether the strong ceh-20 allele ay38(kindly supplied by M. Stern) is likely to be a null mutation, we used direct PCR sequencing to show that it is a C→T transition predicted to cause premature termination of translation (Q102→stop, as also found by E. Chen and M. Stern, personal communication) near the start of the PBC-B domain. Genetic tests in an unc-62(e644) background showed that homozygousceh-20(ay38) results in higher lethality than eitherceh-20(RNAi) or the hemizygous mutant ceh-20(ay38)/nDf16. Moreover, heterozygous ay38/+ caused a dominant enhancement of lethality, and ay38/ay38/+ caused a more severe dominant enhancement. These results indicate that ay38 exhibits antimorphic character in this background and, therefore, is not a null allele (data available on request).

Viability assays

Embryonic and larval lethality were scored by allowing hermaphrodites to lay eggs for 24 hours (20°C) and then transferring the parents to a new plate daily for the next 3-4 days. Twenty-four hours after the hermaphrodite was removed from a plate, all progeny were scored for viability by counting unhatched embryos and living larvae. To assess lethality at later larval stages, surviving animals were followed until they died or became fertile adults.

Microscopy and lineage analysis

Phenotypic analyses and immunofluorescence microscopy were carried out as described (Edgar et al., 2001)using a Leitz microscope equipped with Nomarski optics and UV epi-illumination. Antibodies to LIN-26 and MHC-A were kindly provided by M. Labouesse and R. Waterston, respectively; an integrated jam-1::GFPreporter construct was obtained from M. Han; the ceh-20::GFP reporter construct pHK110, including 2.2 kb of ceh-20 promoter sequence and encoding full-length CEH-20 fused at the C terminus with GFP, was kindly provided by H. Kagoshima and T. Burglin. Lineage analysis was performed with a multi-focal-plane time-lapse video recording system(Thomas et al., 1996); embryos were typically recorded for 5-6 hours post-fertilization at 22°C.

Identification of unc-62 lesions

Genomic DNA fragments were amplified using suitable PCR primers across a∼27 kb region containing unc-62. All unc-62 lesions were identified and confirmed by direct PCR sequencing. For s472, the proximal and distal endpoints are located in the sequences TAAGTTGTCCTTAGTCTTATTAAAA and GCTTTGTGTGTGTGTGAACAGTTAT,respectively (underlined bases are deleted). For ct344, the proximal and distal endpoints are located in the sequences ACTTCAAACGATCAAATAATGATTT and CGGGTTTACAGCCCTGAATAGGTT,respectively. The e917 molecular lesion was identified as a probable inversion by inverse PCR sequencing of religated Sau3AI fragments ofe917 genomic DNA; inversion was confirmed by direct PCR sequencing across the break points. One break point was located in the unc-62region of LGVL (cosmid T08H10), 8.25 kb upstream of the exon 1a start codon(see Fig. 6); the other was found on the right arm of V (included in YAC Y40H4A). For additional characterization of this rearrangement, see supplemental data.

Fig. 6.

Physical map of deletion and inversion breakpoints in the vicinity of theunc-62 promoter, showing numbered regions predicted to include possible regulatory elements (see Discussion). Exon sizes are not to scale.

Fig. 6.

Physical map of deletion and inversion breakpoints in the vicinity of theunc-62 promoter, showing numbered regions predicted to include possible regulatory elements (see Discussion). Exon sizes are not to scale.

RNA interference, RNA blots and quantitative PCR

RNAi was performed essentially as described(Montgomery et al., 1998;Tabara et al., 1998). cDNA templates were either obtained from Y. Kohara (unc-62 andceh-20) or generated by RT-PCR from C. elegans embryonic total RNA, and were transcribed using standard methods prior to annealing to produce double-stranded (ds) RNA. dsRNA was used either for injection(Montgomery et al., 1998) or soaking(http://nematode.lab.nig.ac.jp/db/rnai_s/RNAiBySoaking.htm)(spermidine and gelatin were omitted from the soaking mix) at concentrations of 0.2-1.0 mg/ml. Progeny embryos scored for viability and subsequent larval defects were collected during a 24-96 hour period after injection or a 48-96 hour period after soaking.

For RNA blots total RNAs from the populations described below were isolated, run on formaldehyde gels and blotted onto Hybond-N membranes(Amersham) as described previously (Streit et al., 1999).

For quantitative PCR RNA samples were reverse transcribed with anunc-62 gene-specific primer using SuperScript II™ (Invitrogen)under conditions described by the manufacturer. PCR assays with primer pairs recognizing common sequences near the 3′ and 5′ ends of the mRNAs showed that full-length cDNAs had been generated. Each cDNA sample was then amplified in triplicate with four transcript-specific primer pairs as well as a reference primer pair common to all four transcripts. Amplification reactions were carried out simultaneously in an ABI Prism 7700 Sequence Detection System, using SYBR® Green PCR Master Mix (Applied Biosystems) to allow the course of each reaction to be followed. Using the mean values from each set of three reactions, ratios of each individual transcript to totalunc-62 mRNA were determined. The relative ratios were calculated by the 2▵▵CT method(Livak and Schmittgen, 2001),using the common reference primer pair as the internal control and the EE sample as the calibrator for each primer pair.

GenBank Accession Numbers

Accession numbers for the four unc-62 transcripts 1a-7a, 1a-7b,1b-7a and 1b-7b are AF427474, AF427475, AF427476 and AF427477,respectively.

RESULTS

unc-62 is the C. elegans Meis-family homeobox gene

The unc-62(e644) mutation was mapped to a small interval on the left arm of LGV (Fig. 1A). This region includes the homeobox gene ceh-25, which encodes the onlyC. elegans homolog of Drosophila Hth, a Meis-family member of the TALE class of homeodomain proteins(Burglin, 1997). We established that ceh-25 was unc-62 by identifying molecular lesions associated with this gene for six unc-62 alleles, as described below. We will refer to this gene henceforth as unc-62.

Fig. 1.

unc-62 molecular cloning, transcript structure and sequence comparisons. (A) Genetic and physical maps of the unc-62 region on LGV. Genetic mapping placed unc-62 on the left arm of LGV, between two strain-specific polymorphisms stP3 and RW#L63 and in the region uncovered by deficiency sDf27 but not sDf50. Cosmids mapped to this region by the C. elegans genome consortium are shown below. unc-62 putative regulatory sequences and coding sequence reside on cosmids T08H10 and T28F12, respectively; both the regulatory and coding sequences are included in the fosmid clone H06N16. (B) Structure ofunc-62 transcripts and locations of mutant lesions. unc-62produces four alternatively spliced transcripts that differ in the choice of the first exon (1a or 1b) and the seventh exon, which is the first exon of the homeobox region (7a or 7b). The HM domain is encoded by exons 2, 3, 4 and 5,while the TALE homeodomain is encoded by exons 7a or 7b, 8 and 9. Locations of the point mutations t2012 and e644, the deletionss472 and ct344, and the left inversion breakpoint ofe917 are indicated (see Fig. 6 for a more precise map of the rearrangements). (C,D) Sequence similarities of the conserved UNC-62, Homothorax, and murine Meis1 HM and homeodomains, respectively. Dashes indicate identical residues.

Fig. 1.

unc-62 molecular cloning, transcript structure and sequence comparisons. (A) Genetic and physical maps of the unc-62 region on LGV. Genetic mapping placed unc-62 on the left arm of LGV, between two strain-specific polymorphisms stP3 and RW#L63 and in the region uncovered by deficiency sDf27 but not sDf50. Cosmids mapped to this region by the C. elegans genome consortium are shown below. unc-62 putative regulatory sequences and coding sequence reside on cosmids T08H10 and T28F12, respectively; both the regulatory and coding sequences are included in the fosmid clone H06N16. (B) Structure ofunc-62 transcripts and locations of mutant lesions. unc-62produces four alternatively spliced transcripts that differ in the choice of the first exon (1a or 1b) and the seventh exon, which is the first exon of the homeobox region (7a or 7b). The HM domain is encoded by exons 2, 3, 4 and 5,while the TALE homeodomain is encoded by exons 7a or 7b, 8 and 9. Locations of the point mutations t2012 and e644, the deletionss472 and ct344, and the left inversion breakpoint ofe917 are indicated (see Fig. 6 for a more precise map of the rearrangements). (C,D) Sequence similarities of the conserved UNC-62, Homothorax, and murine Meis1 HM and homeodomains, respectively. Dashes indicate identical residues.

Sequencing of five unc-62 cDNA clones from the C. elegansEST collection, as well as RT-PCR and 5′-RACE products confirmed the general gene structure predicted by Burglin(Burglin, 1997). However, our analysis revealed a second start exon 5′ to that predicted originally,and also showed that the 5′ end of exon 7a is 249 bp downstream of the originally predicted position (Fig. 1B). The 12 exons of unc-62 are distributed over about 15 kb of DNA. The sequencing revealed four alternative transcripts that differ in choice of the first exon (1a or 1b) and the seventh exon (7a or 7b). 5′-RACE analysis showed no evidence for any further upstream exons, but demonstrated that exon 1a is trans-spliced to the splice leader SL1, while exon 1b is not. Exons 7a and 7b encode alternative polypeptide sequences that include the most N-terminal region of the homeodomain. Of the two, the 7a-encoded sequence is more similar to the Drosophila Hth and murine Meis homeodomains. Three of the five cDNA clones sequenced were full length;two of these contained exon 1a and one contained exon 1b. However, all five cDNAs contained exon 7b; transcripts containing exon 7a were detected only by RT-PCR. No transcripts were detected that contained both 7a and 7b. The four alternative transcripts will be referred to subsequently as 1a-7a, 1a-7b,1b-7a and 1b-7b, respectively (see Materials and Methods for GenBank Accession Numbers).

Differential accumulation of unc-62 transcripts

RNA blot analysis of stage-specific total RNA with transcript-specific probes showed different accumulation patterns for the four unc-62mRNAs (Fig. 2A). The 1a-containing mRNAs were more abundant than the 1b-containing mRNAs, but their temporal patterns of accumulation appeared similar. Both were present in adult germline, pregastrulation embryos, later embryos and adult soma. The 7a signals were too weak to assess relative abundances at different stages. The 7b-containing mRNAs appeared much more abundant than the 7a RNAs in both embryos and adults

Fig. 2.

(A) Blots of RNA from early (pregastrulation) embryos (EE), mixed-stage embryos (ME), adult soma (AS) and adult total (AT) hybridized to four transcript-specific probes. EE: >90% pregastrulation embryos(Schauer and Wood, 1990). ME:mixed-stage embryos isolated by hypochlorite treatment ofhim-8(e1489) hermaphrodites. AS: glp-4(bn2ts) young adult hermaphrodites that had been shifted to non-permissive temperature (25°C)at hatching; these animals lack almost all germline cells(Beanan and Strome, 1992). AT:fem-2(b245ts) young adult hermaphrodites that had been shifted to non-permissive temperature (25°C) for 36 hours after hatching (the temperature-sensitive period), then returned to permissive temperature; these animals produce no sperm and therefore contain oocytes but no fertilized embryos (Kimble et al., 1984). Photographs of 18S rRNA UV absorption bands on the blots before probing are shown as loading controls (lower panels). The fourth (AT) lane of each gel is clearly overloaded relative to the other three. In preliminary experiments using equal specific activities for the 1a and 1b probes, the 1b bands were fainter than the 1a bands. In the experiment shown, a 1b probe at about twice the specific activity of the 1a probe was used in order to better compare band intensities among the four RNA preparations. (B) Results of real-time quantitative PCR experiments to compare the relative ratios of each transcript with total unc-62 mRNAs in the same RNA populations as analyzed in A. Figures in the table represent the ratio obtained for each primer pair relative to the ratio obtained for the EE sample, which was arbitrarily assigned a value of 1.0. The ranges shown for each ratio were calculated from the results for each triplicate set according to User Bulletin #2: ABI Prism 7700 Sequence Detection System (2001). This experiment was carried out three times with similar results; ratios and ranges shown in the figure are from one experiment.

Fig. 2.

(A) Blots of RNA from early (pregastrulation) embryos (EE), mixed-stage embryos (ME), adult soma (AS) and adult total (AT) hybridized to four transcript-specific probes. EE: >90% pregastrulation embryos(Schauer and Wood, 1990). ME:mixed-stage embryos isolated by hypochlorite treatment ofhim-8(e1489) hermaphrodites. AS: glp-4(bn2ts) young adult hermaphrodites that had been shifted to non-permissive temperature (25°C)at hatching; these animals lack almost all germline cells(Beanan and Strome, 1992). AT:fem-2(b245ts) young adult hermaphrodites that had been shifted to non-permissive temperature (25°C) for 36 hours after hatching (the temperature-sensitive period), then returned to permissive temperature; these animals produce no sperm and therefore contain oocytes but no fertilized embryos (Kimble et al., 1984). Photographs of 18S rRNA UV absorption bands on the blots before probing are shown as loading controls (lower panels). The fourth (AT) lane of each gel is clearly overloaded relative to the other three. In preliminary experiments using equal specific activities for the 1a and 1b probes, the 1b bands were fainter than the 1a bands. In the experiment shown, a 1b probe at about twice the specific activity of the 1a probe was used in order to better compare band intensities among the four RNA preparations. (B) Results of real-time quantitative PCR experiments to compare the relative ratios of each transcript with total unc-62 mRNAs in the same RNA populations as analyzed in A. Figures in the table represent the ratio obtained for each primer pair relative to the ratio obtained for the EE sample, which was arbitrarily assigned a value of 1.0. The ranges shown for each ratio were calculated from the results for each triplicate set according to User Bulletin #2: ABI Prism 7700 Sequence Detection System (2001). This experiment was carried out three times with similar results; ratios and ranges shown in the figure are from one experiment.

Non-uniformity of sample loading made quantitation from RNA blots difficult(Fig. 2A). Further evidence for differential accumulation was obtained using quantitative real-time PCR analysis (QPCR) with appropriate transcript-specific primers(Giulietti et al., 2001;Livak and Schmittgen, 2001). These experiments allowed us to compare the ratio of each transcript to totalunc-62 mRNA at later stages relative to this ratio in early embryos. The results, summarized in Fig. 2B, are consistent with the RNA blot results, but also show that the levels of the rare 7a transcripts increase over 100-fold from early embryos to adulthood. In addition, the QPCR results show that the levels of 7b mRNAs (unclear from the RNA blot because of non-uniform loading) are approximately the same in adult soma and complete adults, indicating that both 7b and 7a mRNAs are present in adult soma. The lower relative levels of 7a mRNAs in complete adults compared with adult soma raises the possibility that 7a mRNAs are not present in the hermaphrodite germline.

Functional tests for differential transcript requirements during development were attempted using RNAi with four transcript-specific dsRNAs,prior to the publication of evidence for spreading of RNAi that makes such experiments unlikely to be informative(Lipardi et al., 2001). The dsRNAs used had to be small in order to be transcript specific, and did not cause significant phenotypic defects (data not shown).

Molecular lesions in unc-62 mutants

The six mutants analyzed phenotypically (see following section) included a severe zygotic embryonic lethal (s472), two mutants with weaker zygotic effects (e644, ku234) and three primarily maternal-effect lethal mutants of different severities. Sequencing of unc-62 from mutant strains revealed molecular lesions corresponding to all six of these alleles. Three of the mutations affect coding sequence, and two affect upstream putative regulatory sequences(Fig. 1B).

The severe zygotic allele s472 is a 6.3 kb deletion that removes 1.6 kb of upstream sequence, both alternative first exons 1a and 1b, and exons 2, 3 and 4, which encode most of the HM domain. This allele seems likely to be a molecular null. The weaker zygotic allele e644 is a single base transition that changes the Trp410 (TGG) codon to a stop (TAG) codon in the alternatively spliced exon 7b, corresponding to the 19th residue of the homeodomain (Fig. 1D). The resulting transcript should be degraded by the smg system for nonsense-mediated decay (Pulak and Anderson, 1993). A test for effects of inactivating thesmg system on phenotypes resulting from the e644 mutation showed a substantial increase in viability, from 29% (n=968) in the original smg(+) strain to 69% (n=932) in asmg-1 mutant background. Therefore, e644 is acting as a hypomorphic rather than an antimorphic allele. The weakest zygotic alleleku234 is a single base transition that changes the predicted initial Met (ATG) codon in exon 1b to an Ile (ATA) codon. The next predicted Met codon does not occur until the 39th codon of the 1b transcript, 16 codons before the start of the HM-domain coding region in exon 2(Fig. 1C).

The most severe of the maternal-effect lethal mutations, t2012, is a single base transition that changes the first possible translational start in exon 1a from a Met (ATG) to an Ile (ATA) codon. The next predicted Met codon is in exon 2 as indicated above; this is codon 73 of the exon 1a transcript. The less severe maternal-effect lethal mutation ct344 is a 7.2 kb deletion, beginning 7.2 kb upstream of exon 1a. The deletion removes∼3.6 kb of potential unc-62 regulatory sequence, as well as a putative upstream gene (T08H10.1), which is predicted to encode the enzyme aldose reductase. In RNAi experiments, injection of N2 hermaphrodites with dsRNA made using a 0.9 kb cDNA for this gene caused increases of only 5%embryonic and 4% subsequent larval lethality over controls (n=1193 progeny), suggesting that loss of T08H10.1 function in ct344 does not contribute significantly to the resulting phenotype. The third maternal-effect lethal mutation, e917, is a complex chromosomal rearrangement (see Materials and Methods; see supplemental data ). One inversion breakpoint occurs 8.25 kb upstream of exon 1a, within the region deleted by the ct344 mutation, while the other is on the right arm of LGV.

unc-62 mutations result in a variety of phenotypes

The deletion allele s472 causes 100% zygotic lethality among the homozygous progeny of heterozygous hermaphrodites. Of these, 84% die as unhatched embryos, some of which fail to enclose completely and rupture at elongation [Table 1;Fig. 3A; supplemental Fig. S2]. The remaining 16% survive to hatching but display a variety of morphological defects consistent with failures of hypodermal and muscle patterning and differentiation. Lineage analysis (n=3) showed misaligned cleavages and mispositioning of the C (dorsal) and V (lateral seam cell) precursors as early as the 100-cell stage, resulting in V cells in dorsal positions and C cells more lateral and posterior than in wild-type embryos, as described below for the two strong maternal-effect alleles(Fig. 4B). C muscle precursors were also mispositioned (n=1). Ventral cleft closure, the last stage of gastrulation (at ∼300 cells), was incomplete at the posterior in two out of three embryos lineaged. In three out of three, subsequent hypodermal enclosure was delayed and a few cells were extruded [supplemental Fig. S2 at(Supplementary Information)]. These embryos failed to elongate. Mutant s472 embryos expressing a JAM-1::GFP reporter, which localizes to the tight junctions between hypodermal cells, showed variable hypodermal cell defects: seam cells could be reduced in number, not touching or grouped in patches(Fig. 4G,H). In contrast tonob-1 embryos (Van Auken et al.,2000), the gut (E) lineage appeared normal (n=1).

Table 1.

Zygotic and maternal-effect lethality resulting from six unc-62alleles*

Viability among total progeny
Lethality among unc-62 homozygous mutant progeny(%)
RowMaternal genotypeEmbryonic arrest(%)Larval arrest(%)Surviving progeny(%)
\(\frac{+unc-62(s472)unc-46(e177)+}{unc-60(e723)++dpy-11(e224)}\)
 
21 75 100 
\(\frac{unc-62(e644)dpy-11(e224)}{++}\)
 
88 20 
\(\frac{unc-62(e644)}{unc-62(e644)}\)
§ 
18 49 33 67 
\(\frac{unc-62(t2012)dpy-11(e224)}{++}\)
 
89 
\(\frac{unc-62(t2012)dpy-11(e224)}{unc-62(t2012)dpy-11(e224)}\)
 
96 100 
\(\frac{unc-62(ct344)dpy-11(e224)}{++}\)
 
91 
\(\frac{unc-62(ct344)dpy-11(e224)}{unc-62(ct344)dpy-11(e224)}\)
 
57 43 100 
\(\frac{unc-62(e917)}{+}\)
 
95 1-4 
\(\frac{unc-62(e917)}{unc-62(e917)}\)
 
57 40 97 
10 
\(\frac{unc-62(ku234)}{unc-62(ku234)}\)
 
87 13 
11 
\(\frac{+}{+}\)
 
96 
12 
\(\frac{unc-46(e177)}{unc-46(e177)}\)
 
96 
13 
\(\frac{+unc-46(e177)}{unc-60(e273)dpy-11(e224)}\)
 
ND ND 
Viability among total progeny
Lethality among unc-62 homozygous mutant progeny(%)
RowMaternal genotypeEmbryonic arrest(%)Larval arrest(%)Surviving progeny(%)
\(\frac{+unc-62(s472)unc-46(e177)+}{unc-60(e723)++dpy-11(e224)}\)
 
21 75 100 
\(\frac{unc-62(e644)dpy-11(e224)}{++}\)
 
88 20 
\(\frac{unc-62(e644)}{unc-62(e644)}\)
§ 
18 49 33 67 
\(\frac{unc-62(t2012)dpy-11(e224)}{++}\)
 
89 
\(\frac{unc-62(t2012)dpy-11(e224)}{unc-62(t2012)dpy-11(e224)}\)
 
96 100 
\(\frac{unc-62(ct344)dpy-11(e224)}{++}\)
 
91 
\(\frac{unc-62(ct344)dpy-11(e224)}{unc-62(ct344)dpy-11(e224)}\)
 
57 43 100 
\(\frac{unc-62(e917)}{+}\)
 
95 1-4 
\(\frac{unc-62(e917)}{unc-62(e917)}\)
 
57 40 97 
10 
\(\frac{unc-62(ku234)}{unc-62(ku234)}\)
 
87 13 
11 
\(\frac{+}{+}\)
 
96 
12 
\(\frac{unc-46(e177)}{unc-46(e177)}\)
 
96 
13 
\(\frac{+unc-46(e177)}{unc-60(e273)dpy-11(e224)}\)
 
ND ND 
*

Individual embryos from the indicated parental genotypes were picked and scored for hatching at 20°C. Inviable embryos were subtracted from animals that grew to adulthood to determine larval lethality. Between 1000 and 2000 embryos were scored for each genotype.

The genotype-independent background lethality of 2% embryonic and 2% larval seen in controls (rows 11-13) can be subtracted from lethality and added to viability figures in rows 1-12 to estimate the effects resulting from the genotypes shown. Surviving progeny were defined as those that grew to adulthood.

Percent lethality among homozygous unc-62 mutant progeny of heterozygotes was estimated as follows for the various genotypes. For row 1,no Unc-46 progeny survived; therefore, homozygotes were 100% inviable. For rows 2, 4 and 6 the percentages of Dpy survivors among total progeny were 19%,22% and 22%, respectively. If Dpys, which are assumed to represent 25% of progeny, die at the control frequency of 4%, then the expected percentage of Dpy survivors among total progeny would be 24%. The discrepancy between this and the observed value represents the missing Dpy animals, and can be divided by 25% to give the percentage of lethality among homozygous mutant progeny. Thus in row 2, for example, the discrepancy is 5% (24% versus 19%), which indicates 5/25 or 20% lethality among homozygotes. For row 8, these calculations could not be made as no marker was present. Thus, 1% is the minimum homozygote lethality if homozygotes were at no disadvantage; 4% is the maximum if only homozygotes failed to survive.

§

A similar experiment with unc-62(e644) dpy-11(e224) gave similar results (Table 3, row 7).

Fig. 3.

Late embryonic and larval arrest phenotypes resulting from fiveunc-62 alleles. Animals in row A are homozygous mutant progeny of heterozygous unc-62/+ hermaphrodites; those in rows B-E are progeny of homozygous mutant hermaphrodites. Alleles are indicated in each panel. Each allele results in a range of phenotypes represented by the individuals shown.

Fig. 3.

Late embryonic and larval arrest phenotypes resulting from fiveunc-62 alleles. Animals in row A are homozygous mutant progeny of heterozygous unc-62/+ hermaphrodites; those in rows B-E are progeny of homozygous mutant hermaphrodites. Alleles are indicated in each panel. Each allele results in a range of phenotypes represented by the individuals shown.

Fig. 4.

Aberrant organization of early embryonic dorsal hypodermal and seam cell precursors and the resulting differentiated cells in unc-62 mutant embryos. Scale bars represent ∼10 μm. (A-D) Nomarski images of embryos(dorsal or ventral view, dorsal focal plane, anterior leftwards) that were lineaged to identify cells and photographed about 4 hours post-fertilization(22-25°C). Precursors of hypodermal cells derived from the C lineage and seam cells derived from V lineages are positioned as indicated. (A) Wild-type embryo. Note that dorsal C-derived hypodermal precursors are flanked by V-lineage cells. (Although this embryo is shown approximately one division earlier than the other embryos, the relative position of dorsal midline C hypodermal cells and lateral V precursors is already established at this time,and does not change.) (B) s472 mutant embryo. Some of the C hypodermal precursors lie abnormally ventral to this focal plane, and are not visible. (C) t2012 mutant embryo. (D) ct344 mutant embryo. Embryos in B-D are about one division later than the embryo in A. Relative positions of the C- and V-derived hypodermal cells are aberrant in the three mutants. (E-N) Nomarski and fluorescent images (lateral views, surface focal plane, anterior leftwards) of embryos and larvae expressing JAM-1::GFP, a marker for hypodermal cell junctions. (E,F) Wild-type newly hatched L1,showing the normal expression of JAM-1 at the borders of the row of seam cells. The dorsal and ventral hypodermal cells have fused into syncytia,showing no cell boundaries. (G,H) s472 newly hatched L1. There are few outlined cells, which do not show a contiguous seam cell pattern. (I,J)e644 L1, with a more normal pattern but some dorsal hypodermal cells that have not fused. (K,L) t2012 newly hatched L1, with disconnected seam cells. (M,N) ct344 late embryo, with few, abnormally large seam cells.

Fig. 4.

Aberrant organization of early embryonic dorsal hypodermal and seam cell precursors and the resulting differentiated cells in unc-62 mutant embryos. Scale bars represent ∼10 μm. (A-D) Nomarski images of embryos(dorsal or ventral view, dorsal focal plane, anterior leftwards) that were lineaged to identify cells and photographed about 4 hours post-fertilization(22-25°C). Precursors of hypodermal cells derived from the C lineage and seam cells derived from V lineages are positioned as indicated. (A) Wild-type embryo. Note that dorsal C-derived hypodermal precursors are flanked by V-lineage cells. (Although this embryo is shown approximately one division earlier than the other embryos, the relative position of dorsal midline C hypodermal cells and lateral V precursors is already established at this time,and does not change.) (B) s472 mutant embryo. Some of the C hypodermal precursors lie abnormally ventral to this focal plane, and are not visible. (C) t2012 mutant embryo. (D) ct344 mutant embryo. Embryos in B-D are about one division later than the embryo in A. Relative positions of the C- and V-derived hypodermal cells are aberrant in the three mutants. (E-N) Nomarski and fluorescent images (lateral views, surface focal plane, anterior leftwards) of embryos and larvae expressing JAM-1::GFP, a marker for hypodermal cell junctions. (E,F) Wild-type newly hatched L1,showing the normal expression of JAM-1 at the borders of the row of seam cells. The dorsal and ventral hypodermal cells have fused into syncytia,showing no cell boundaries. (G,H) s472 newly hatched L1. There are few outlined cells, which do not show a contiguous seam cell pattern. (I,J)e644 L1, with a more normal pattern but some dorsal hypodermal cells that have not fused. (K,L) t2012 newly hatched L1, with disconnected seam cells. (M,N) ct344 late embryo, with few, abnormally large seam cells.

The canonical allele e644, in addition to the originally described Unc phenotype (Brenner, 1974),results in zygotic defects that cause 20% lethality among the homozygous mutant progeny of heterozygous hermaphrodites, primarily during larval stages(Table 1). The e644homozygotes that survive to adulthood exhibit a much higher level (67%) of progeny lethality (Table 1),indicating a significant maternal effect. Some of these progeny die as embryos, but about half are inviable larvae with variable phenotypes, some of which probably result from abnormal embryonic development(Fig. 3B). The defective embryos show variable misplacement of seam cells and failure of hypodermal cells to fuse normally into hypodermal syncytia(Fig. 4I,J). Survivinge644 mutant adults display several phenotypes, including Unc,egg-laying-defective (Egl) and variably abnormal larval morphology (Vab). Previous studies suggested that the Unc movement defects may be a result of aberrant axonal outgrowth of motoneurons(Siddiqui, 1990). The Egl animals appear to have normal vulval structures but fail to lay eggs, even when treated with serotonin (5-HT) (data not shown), which stimulates the vulval muscles to promote egg-laying(Thomas et al., 1990;Trent et al., 1983). This result suggests that the Egl animals are defective in vulval muscles rather than vulval innervation. Causes of the Vab phenotype are unknown.

The third non-maternal allele, ku234, was isolated in independent screens for Egl mutations (M. S. and M. H., unpublished). Although 13% of the progeny of homozygous ku234 hermaphrodites die as embryos and larvae(Table 1), the survivors exhibit no gross morphological defects during embryonic or larval stages. However, about 50% of the survivors have visible vulval defects(Fig. 5) and 100% are Egl. The vulval phenotypes of ku234/s472 and the hemizygousku234/sDf27 are only marginally more severe than those ofku234 homozygotes (Fig. 5). The ku234 mutation was identified as anunc-62 allele on the basis of its failure to complement s472and e644 (see Materials and Methods and supplemental data ). However, it does complement the maternal-effect lethal alleles t2012,ct344 and e917.

Fig. 5.

Examples of aberrant vulval development in ku234 mutant L4 hermaphrodites. (A) Wild type. (B) ku234: descendants of P7.p (arrow)have not joined in the main invagination. About 50% of ku234 exhibit this phenotype (non-Vul-abnormal), while the rest have apparently normal vulvae. (C) ku234/s472: descendents of P7.p (arrow) appear non-vulval. About a third of the non-Dpy segregants from ku234 dpy-11/unc-62(s472) unc-46 exhibited this semi-vulvaless phenotype, which is only slightly more severe that the non-Vul-abnormal phenotype in B. The remaining non-Dpy animals were either non-Vul-abnormal or normal. (D)ku234/sDf27: posterior Pn.p cells appear to be generating vulval-like descendants near the tail (arrow). This phenotype was seen only occasionally. As in C, about a third of the non-Dpy segregants of ku234 dpy-11/sDf27 unc-46 exhibited the semi-vulvaless phenotype, while the rest were non-Vul-abnormal or normal.

Fig. 5.

Examples of aberrant vulval development in ku234 mutant L4 hermaphrodites. (A) Wild type. (B) ku234: descendants of P7.p (arrow)have not joined in the main invagination. About 50% of ku234 exhibit this phenotype (non-Vul-abnormal), while the rest have apparently normal vulvae. (C) ku234/s472: descendents of P7.p (arrow) appear non-vulval. About a third of the non-Dpy segregants from ku234 dpy-11/unc-62(s472) unc-46 exhibited this semi-vulvaless phenotype, which is only slightly more severe that the non-Vul-abnormal phenotype in B. The remaining non-Dpy animals were either non-Vul-abnormal or normal. (D)ku234/sDf27: posterior Pn.p cells appear to be generating vulval-like descendants near the tail (arrow). This phenotype was seen only occasionally. As in C, about a third of the non-Dpy segregants of ku234 dpy-11/sDf27 unc-46 exhibited the semi-vulvaless phenotype, while the rest were non-Vul-abnormal or normal.

The three alleles t2012, ct344 and e917 result primarily in maternal-effect lethality. Although they also exhibit zygotic effects to different degrees, as discussed below, most of the homozygous progeny embryos from hermaphrodites heterozygous for these alleles develop into fertile adults that appear normal (Table 1). These homozygous hermaphrodites produce nearly 100% inviable embryos and defective larvae by self-fertilization. However, when mated to wild-type males, both t2012 and ct344 hermaphrodites produce 80-90%viable progeny of genotype unc-62/+(Table 2). Similar rescue experiments with unc-62/+ males indicate that rescue depends on introduction of the + allele (i.e. there is no paternal effect; all surviving progeny are of genotype unc-62/+). e917 hermaphrodites exhibit behavior similar to t2012 and ct344 hermaphrodites in both types of experiments (data not shown). Therefore, these three alleles behave as partial (male-rescuable) maternal rather than strict maternal-effect mutations.

Table 2.

Lethality resulting from heteroallelic combinations of unc-62alleles*

RowMaternal genotypePaternal genotypeEmbryonic arrest (%)Larval arrest (%)Viable (%)n
\(\frac{e644}{e644}\)
 
self NA 37 54 352 
\(\frac{e644}{e644}\)
 
× ++ 98 1093 
\(\frac{e644}{e644}\)
 
× ct344 dpy-11§ 12 29 59 298 
\(\frac{e644}{e644}\)
 
× t2012 dpy-11§ 20 27 53 202 
\(\frac{e644+}{ct344{\ }dpy-11}\)
 
self NA 16 31 53 794 
\(\frac{e644+}{t2012{\ }dpy-011}\)
 
self NA 10 62 28 677 
ct344 dpy-101 self NA 57 42 <1 1353 
ct344 dpy-11 × + + 13 86 503 
ct344 dpy-11 × e644§ 14 21 64 285 
10 ct344 dpy-11 × t2012 dpy-11§ 41 59 105 
11 
\(\frac{ct344{\ }dpy-11}{t2012{\ }dpy-11}\)
 
self NA 90 10 336 
12 t2012 dpy-11 self NA 96 534 
13 t2012 dpy-11 × + + 10 89 1000 
14 t2012 dpy-11 × ct344 dpy-11§ 64 36 173 
15 t2012 dpy-1e644§ 76 24 165 
16 self NA <1 <1 >99 >1000 
17 × + + 94 474 
RowMaternal genotypePaternal genotypeEmbryonic arrest (%)Larval arrest (%)Viable (%)n
\(\frac{e644}{e644}\)
 
self NA 37 54 352 
\(\frac{e644}{e644}\)
 
× ++ 98 1093 
\(\frac{e644}{e644}\)
 
× ct344 dpy-11§ 12 29 59 298 
\(\frac{e644}{e644}\)
 
× t2012 dpy-11§ 20 27 53 202 
\(\frac{e644+}{ct344{\ }dpy-11}\)
 
self NA 16 31 53 794 
\(\frac{e644+}{t2012{\ }dpy-011}\)
 
self NA 10 62 28 677 
ct344 dpy-101 self NA 57 42 <1 1353 
ct344 dpy-11 × + + 13 86 503 
ct344 dpy-11 × e644§ 14 21 64 285 
10 ct344 dpy-11 × t2012 dpy-11§ 41 59 105 
11 
\(\frac{ct344{\ }dpy-11}{t2012{\ }dpy-11}\)
 
self NA 90 10 336 
12 t2012 dpy-11 self NA 96 534 
13 t2012 dpy-11 × + + 10 89 1000 
14 t2012 dpy-11 × ct344 dpy-11§ 64 36 173 
15 t2012 dpy-1e644§ 76 24 165 
16 self NA <1 <1 >99 >1000 
17 × + + 94 474 
*

Experiments in this table were carried out independently by a different investigator from those in Table 1. Discrepancies in comparable results between the two tables are at least partly due to differences in scoring methods and criteria.

Not applicable.

Hermaphrodites of the indicated maternal genotype were mated with males of the indicated paternal genotype.

§

Male was heterozygous [unc-62(paternal)/+]. The unc-62heteroallelic mutant cross progeny could be distinguished as Dpy in crosses between dpy-11 hermaphrodites and dpy-11/+ males, and as Unc progeny in crosses involving e644 and either t2012 orct344, because e644/t2012 and e644/ct344 animals exhibit the Unc-62 phenotype. Wild-type cross progeny[unc-62(maternal)/+] were not included in these counts and all embryonic lethality was attributed to theunc-62(maternal)/unc-62(paternal) genotype.

Hermaphrodites carried the fem-1(hc17ts) mutation and were reared at 25°C to cause self-sterility. Males were of genotypehim-8(e1489).

For t2012, the most severe of the maternal-effect alleles, almost all the self-progeny of homozygous hermaphrodites arrest as unhatched embryos(Table 1). These embryos execute the normal embryonic lineage pattern in terms of timing and numbers of divisions (n=2) and produce differentiated gut, muscle and pharyngeal tissue. However, they exhibit early misplacement of hypodermal precursor cells(Fig. 4C) and generally fail to undergo normal enclosure, elongation and morphogenesis(Fig. 3C). In addition, they fail to produce the normal number of differentiated hypodermal cells in embryos and larvae as identified using antibodies to the hypodermal marker LIN-26 (Labouesse et al.,1996). LIN-26, normally expressed in ∼84 cells, was detected(by immunofluorescence using an anti-LIN-26 antibody) in an average of only 64 cells (range 53-75; n=28) among progeny of homozygous t2012hermaphrodites. Severe disorganization was seen among these cells int2012 mutant embryos expressing the JAM-1::GFP reporter construct(Fig. 4K,L), as well as among body-wall muscle cells visualized with anti-MHC-A (not shown). Thet2012 allele also results in some zygotic lethality(Table 1) with similar gross embryonic phenotypes.

Homozygous ct344 hermaphrodites produce embryos that generally arrest later than those from t2012 animals, resulting in roughly equal numbers of unhatched embryos and hatched inviable larvae(Table 1). Embryos usually progress through morphogenesis, but often display a Nob phenotype in which anterior head and pharyngeal structures develop normally, but the posterior is severely malformed (Fig. 3D). As in t2012 embryos, although the embryonic cell lineage pattern ofct344 embryos appears normal (n=2), hypodermal precursors are misplaced by the early gastrulation stage(Fig. 4D), and hypodermal cells are under-represented and grossly disorganized in Nob larvae, as if they fail to differentiate properly (Fig. 4M,N). The ct344 allele also results in some zygotic lethality (Table 1) with similar gross embryonic phenotypes.

Table 2 shows the lethality resulting from heteroallelic genotypes derived from e644, t2012 andct344. Heteroallelic ct344/t2012 embryos are all inviable,regardless of the parental origins of the two alleles. Interestingly, however,t2012/e644 embryos are >50% viable if t2012 is introduced from the male parent, but 0% viable if the maternal parent is homozygous fort2012. Such non-reciprocality is not observed for ct344/e644embryos; their viability is similar to that of e644 embryos regardless of the parental origin of ct344. These results are consistent with other observations that ct344 is a weaker allele thant2012.

The third maternal-effect allele e917 is similar to ct344in the percentages of embryonic and larval lethality among the self progeny of homozygous hermaphrodites (Table 1). However, the resulting larvae appear to be generally less defective morphologically (Fig. 3E). The ∼3% that survive to be fertile adults display a variety of defects including Egl, Unc and Vab. This allele causes almost no zygotic lethality (Table 1),but the homozygous progeny of heterozygous hermaphrodites display the Egl phenotype. Because e917 is a complex rearrangement as described above and could affect other genes besides unc-62, the resulting phenotypes were not analyzed in detail.

The C. elegans PBC-family genes ceh-20 andceh-40 may function redundantly during embryogenesis

To test whether the strong interactions observed between MEIS- and PBC-family proteins in both Drosophila and vertebrates (see Introduction) might also occur in C. elegans, we first analyzed loss of function phenotypes for the C. elegans PBC-class geneceh-20. Preliminary characterization of the strong ceh-20allele ay38 indicated that it is probably antimorphic (see Materials and Methods) and thus inappropriate for interaction studies. A previous description by Liu and Fire (Liu and Fire,2000) of the ceh-20(RNAi) phenotype, which they concluded to be null, reported failure of post-embryonic development in the M lineage but little or no embryonic lethality. We obtained similar results(Table 3, rows 4 and 5),suggesting that ceh-20 is not essential during embryogenesis.

Table 3.

Interactions between unc-62, ceh-20 and ceh-40

RowGenotype*Embryonic lethality (%)Larval lethality(%)Viable (%)n
N2 control (H2O, I98 1397 
N2 control (H2O, S99 134 
ceh-40(RNAi, S) 96 86 
ceh-20(RNAi, I) 32 66 927 
ceh-20(RNAi, S) 37 59 249 
ceh-20(RNAi, S); ceh-40(RNAi, S) 21 76 231 
unc-62(e644); dpy-11(e224) 14 47 39 378 
unc-62(e644); ceh-40(RNAi, S) 20 55 25 201 
ceh-20(RNAi, I); unc-62(e644) 26 70 544 
10 ceh-20(RNAi, S); unc-62(e644); ceh-40(RNAi,S) 68 28§ 1203 
11 unc-62(RNAi,I) 72 21 1036 
RowGenotype*Embryonic lethality (%)Larval lethality(%)Viable (%)n
N2 control (H2O, I98 1397 
N2 control (H2O, S99 134 
ceh-40(RNAi, S) 96 86 
ceh-20(RNAi, I) 32 66 927 
ceh-20(RNAi, S) 37 59 249 
ceh-20(RNAi, S); ceh-40(RNAi, S) 21 76 231 
unc-62(e644); dpy-11(e224) 14 47 39 378 
unc-62(e644); ceh-40(RNAi, S) 20 55 25 201 
ceh-20(RNAi, I); unc-62(e644) 26 70 544 
10 ceh-20(RNAi, S); unc-62(e644); ceh-40(RNAi,S) 68 28§ 1203 
11 unc-62(RNAi,I) 72 21 1036 
*

All RNA injections were into N2 where no other genotype is indicated.RNAi, I and RNAi, S refer to RNAi by injection and soaking,respectively (see Materials and Methods).

Includes all inviable larvae, ranging from arrest at hatching to lethality at later stages.

Of the 16 injected animals that laid eggs, one produced progeny that were apparently unaffected by injection and were undistinguishable fromunc-62(e644) progeny; the progeny of this animal were excluded from these results.

§

All larvae arrested at hatching and were severely deformed, similarly tounc-62(s472) progeny.

Carried out with dsRNA corresponding to 1a-7b mRNA.

However, there is a second PBC-family homolog in C. elegans, the predicted gene ceh-40. The predicted protein products of the two genes are 41% identical overall and 74% identical in the homeodomain(Burglin, 1997), suggesting that ceh-20 and ceh-40 might function redundantly. Noceh-40 mutations have been described, and no ceh-40 cDNAs were found in current databases. However, microarray experiments have shown that ceh-40 transcripts, absent maternally, appear in embryos at low levels around the onset of gastrulation (L. R. Baugh and C. P. Hunter,personal communication). A ceh-40 cDNA was obtained using RT-PCR with embryonic mRNA, allowing us to confirm the splice sites of the exons predicted from genomic sequence. When ceh-40(RNAi) was carried out in N2 animals using dsRNA made from this template, no significant embryonic or early larval lethality was observed (Table 3, row 3). Consistent with redundant function, the combination ofceh-40(RNAi) with ceh-20(RNAi) caused 21% embryonic lethality (Table 3, row 6), a significant increase over either RNAi alone. However, the phenotypes of the dying embryos were not obviously similar to those resulting from theunc-62(s472) null allele. [While this paper was in revision, a similar identification of ceh-40 was reported by Takacs-Vellai et al.(Takacs-Vellai et al., 2002),who also demonstrated embryonic expression of a ceh-40::GFPconstruct.]

Losses of ceh-20 and ceh-40 functions interact strongly with the unc-62(e644) mutation to give anunc-62(null) phenotype

To test for genetic interactions between the PBC-family genes andunc-62, we used the semi-viable unc-62(e644) allele(Table 1). Because the variety of abnormal phenotypes observed among unc-62(e644) homozygous larvae(Fig. 3) made it impractical to reliably detect enhancement of larval defects, we scored only enhancement of lethality as a measure of ceh-20(RNAi) and ceh-40(RNAi)interaction. The unc-62(e644); ceh-40(RNAi) combination resulted in a slight enhancement of embryonic lethality and overall lethality compared tounc-62(e644) alone (Table 3, rows 7 and 8). The ceh-20(RNAi); unc-62(e644)combination caused an enhancement of embryonic lethality and a larger increase in overall lethality (Table 3,rows 7 and 9). Most strikingly, the triple combination ceh-20(RNAi);unc-62(e644); ceh-40(RNAi) caused high embryonic and overall lethality,similar to that seen with unc-62(RNAi)(Table 3, rows 10 and 11). Moreover, the phenotypes of the dying, hatched larvae resembled those resulting from the null allele unc-62(s472)(Table 1,Fig. 3) andunc-62(RNAi) (Table 3). Other experiments showed that overexpression ofceh-20 enhanced the lethality and other phenotypes resulting froms472/+, e644, and ct344 [see legend to supplemental Fig. S4].

Additional evidence for interaction was obtained in experiments to test whether nuclear localization of a CEH-20::GFP fusion protein would be affected by unc-62 defects, as predicted from studies cited in the Introduction showing that Drosophila Hth and vertebrate Meis functions are required for nuclear localization of Exd and Pbx proteins,respectively. In wild-type embryos, CEH-20::GFP was found to be expressed in many cells, all of which exhibited nuclear localization. In unc-62(ct344),unc-62(e644) and unc-62(RNAi) embryos, expression was still widespread, but nuclear localization, like the other mutant phenotypes, was variable, observed in none to about half of the expressing cells [see supplemental Fig. S4].

DISCUSSION

Maternal and zygotic functions of unc-62 play roles in embryogenesis

We have established that the locus defined genetically as unc-62is the sole C. elegans Meis-class homeobox gene, previously designated ceh-25 (Burglin,1997). The range of embryonic and larval phenotypes resulting fromunc-62 mutations indicates that unc-62 serves multiple functions in C. elegans development, as does its Drosophilahomolog hth (Kurant et al.,2001). We have confirmed, with one correction, the presence of four predicted alternatively spliced transcripts, and have shown that all are present in both embryos and adults. The molecular analysis suggests thats472 is null and prevents synthesis of any functional unc-62transcripts, and that t2012, ku234 and e644 prevent or reduce the translation of functional proteins from 1a-, 1b- and 7b-containing transcripts, respectively (see Fig. 1).

Assuming these suggestions are valid, we can make several predictions about the time of synthesis and function of the unc-62 mRNAs and their translation products. 1a transcripts must be required in the early embryo because the self-progeny of t2012/t2012 hermaphrodites (lacking 1a)are all inviable. Apparently either maternal or zygotic provision of the 1a-containing transcripts is equally effective in promoting survival, because either the self-progeny t2012/t2012 embryos from t2012/+hermaphrodites or the t2012/+ cross-progeny embryos fromt2012/t2012 hermaphrodites mated to wild-type males are ∼90%viable. 1a-containing transcripts are usually sufficient for survival because 87% of the self progeny of ku234/ku234 hermaphrodites (lacking 1b)are viable. However, we observe a strong dependence of viability on dose of 1a transcripts. 1a transcripts can be reduced by 75% (t2012 homozygous progeny of a t2012/+ heterozygote) and animals are still 90% viable,but a further 25% reduction (t2012 homozygous progeny of at2012 homozygote) results in 100% lethality.

We can infer less about time of synthesis or function of 1b-containing transcripts, which are less abundant than 1a transcripts. As mentioned above,1b transcripts may be nonessential for viability, as the ku234 allele causes only 13% lethality. However, their presence can affect viability when levels of 1a transcripts are reduced, based on the difference in viability between t2012 embryos from t2012/+ hermaphrodites ands472 embryos from s472/+ hermaphrodites. Both these embryos should have only 25% the wild-type level of 1a transcripts, yet the former,which can produce normal levels of 1b transcripts, are 90% viable, while the latter, which can produce no 1b transcripts, are <1% viable. There may be additional differences in transcript levels if the ku234 andt2012 missense alleles produce some functional transcripts that thes472 deletion cannot.

The fact that dose effects may contribute to the phenotypes observed does not eliminate the likelihood of transcript-specific functions. This is demonstrated by the 100% penetrance of the Egl phenotype caused by the weakku234 allele when compared with the sporadic appearance of this phenotype resulting from the stronger e644 allele.

With regard to the alternative homeodomains encoded by 7a- and 7b-containing transcripts, survival of up to 33% of the self-progeny ofe644/e644 hermaphrodites (deficient in translation products of 7b-containing mRNAs) suggests that the rarer 7a products may compensate for the lack of 7b products. We cannot be confident that 7a products are present in the early embryo, because our quantitative PCR data suggested that 7a-containing transcripts might not be produced maternally(Fig. 2). Alternatively, other proteins with related functions might compensate for 7b products (see discussion of ceh-20 and ceh-40 below). Interestingly,molecular analysis of unc-62 mRNAs in the related nematode C. briggsae (estimated time of divergence 15-30 Mya; see supplemental Fig. S3) showed that the splice acceptor sequence for exon 7a appears to be missing, and only transcripts corresponding to 7b could be detected (B. R. and W. B. W., unpublished). This finding supports the suggestion that the 7a and 7b derived proteins in C. elegans may have overlapping functions.

The maternal-effect unc-62 alleles define putative upstream regulatory elements

The t2012, ct344, and e917 mutations appear to comprise an allelic series of decreasing severity. All result in maternal-effect lethality that can be rescued by zygotic expression of a wild-type copy introduced from a mating male. The ct344 allele is less severe thant2012, based on generally later arrest of mutant embryos and the finding that progeny of ct344 hermaphrodites can be rescued by mating to e644 males, while progeny of t2012 hermaphrodites can be rescued only by mating to wild-type males. The e917 allele is even less severe, as indicated by the lack of any zygotic lethality and the observation that 3% of the self progeny from homozygous hermaphrodites survive to reproduce.

Based on the similarity of the phenotypes resulting from these three alleles, they seem likely to affect common functions. Because t2012affects translation of la-containing mRNAs, while ct344 ande917 affect a region ∼7 kb upstream of the unc-62-coding sequence (Fig. 6), we postulate that ct344 and e917 disrupt or remove upstream regulatory elements required for the production of la-containing transcripts. Because thee917 inversion results in a very similar maternal-effect phenotype toct344, at least one such enhancer element may lie between thee917 breakpoint and the promoter-distal end of the region deleted byct344 (Fig. 6, region 1). The zygotic lethality caused by ct344, which is not seen withe917, suggests that an additional element may lie in the ∼1 kb region between the e917 inversion breakpoint and the proximal end of the ct344 deletion (Fig. 6, region 2). Additional enhancers may also be present in regions 3 and 4.

The Meis family gene unc-62 and the PBC-family genesceh-20 and ceh-40 interact genetically

In Drosophila, similar phenotypes result from loss of function mutations in the Meis-family homolog hth and the PBC-family homologexd, consistent with other evidence for strong interactions between Meis- and PBC-family proteins (see Introduction). Although inactivation of the only previously studied C. elegans exd homolog, ceh-20,causes only larval and adult defects and no embryonic lethality(Liu and Fire, 2000) (our results above), we have shown that this result may be at least partially due to overlapping functions of two C. elegans PBC-family genes,ceh-20 and ceh-40. These two genes interact genetically:ceh-40(RNAi) causes no significant embryonic lethality, whileceh-20(RNAi); ceh-40(RNAi) causes a substantial decrease in larval lethality with a corresponding increase in embryonic lethality compared withceh-20(RNAi) alone. In addition, whereas unc-62(RNAi)embryos arrest with the Nob phenotype, ceh-20(RNAi); ceh-40(RNAi)embryos arrest earlier, prior to morphogenesis. Recent phylogenetic comparisons with similar genes in other nematodes support the view thatceh-40 may be a C. elegans exd ortholog (A. Aboobaker,personal communication). The RNAi results suggest that ceh-20 andceh-40 functions may overlap, with ceh-40 normally functioning earlier than ceh-20. The incomplete penetrance of theceh-20(RNAi); ceh-40(RNAi) lethality compared with the complete penetrance of unc-62(s472) or unc-62(RNAi) lethality could indicate either that silencing by RNAi was incomplete, or that the functions of ceh-20 and ceh-40 are not essential for unc-62function as exd is essential for hth function inDrosophila.

In the background of the hypomorphic unc-62(e644) allele, which is predicted to reduce or eliminate production of a functional protein from 7b-containing transcripts, ceh-20 and ceh-40 show a pattern of interaction with each other similar to that described above. Individually,their RNAi phenotypes are roughly additive to the unc-62(e644)phenotype. However, the triple combination of ceh-20(RNAi); unc-62(e644);ceh-40(RNAi) results in complete embryonic inviability, with terminal phenotypes similar to those caused by the null allele unc-62(s472). One interpretation of these results is that when levels of the putatively interacting UNC-62 and CEH-20/CEH-40 proteins both fall below some threshold level, the result is a strong Unc-62 phenotype. A more intriguing possibility is that there is some overlap in the functions of CEH-20/CEH-40 and the UNC-62 proteins encoded by 7b-containing transcripts. Perhaps, for example, the latter could act directly as a Hox transcription cofactor.

Possible functions of UNC-62 proteins in embryogenesis

Is UNC-62 a co-factor for the posterior-group homeodomain transcription factors NOB-1 and PHP-3 in embryonic development, as might be expected from findings in other organisms? Table 4 summarizes characteristics of Nob phenotypes resulting from strong lf mutations in unc-62 and nob-1/php-3 as well as, for comparison, the caudal homolog pal-1. Although some aspects of the unc-62 and nob-1 Nob phenotypes are similar, others are not. For example, we did not detect posterior to anterior cell fate transformations in the E lineage of unc-62(s472) embryos,and the severe enclosure defects and hypodermal cell disorganization in these embryos are more similar to pal-1 than to nob-1 Nobs [seeFig. 4 and supplemental Fig. S2 and Fig. S5]. In experiments with a rescuing nob-1::GFP reporter construct, the number of nob-1-expressing cells was not significantly altered inunc-62(RNAi) embryos, and overexpression of nob-1 did not rescue the unc-62 phenotype (E. Kress and T. S., unpublished). Furthermore, presence of the unc-62(e644) allele did not increase the lethality resulting from a weak nob-1 allele (E. Kress and T. S.,unpublished). We conclude that there is so far no convincing evidence for an essential co-factor role, and that the unc-62 phenotypes are not primarily the result of altered NOB-1 function. We have not tested for possible interactions of unc-62 products with the later acting Hox gene functions of lin-39, mab-5 and egl-5. Possible interactions of unc-62 and pal-1 functions are under investigation.

Table 4.

Comparison of Nob embryonic phenotypes resulting from strong lf alleles of unc-62, pal-1 andnob-1*

Defective gene
Phenotypeunc-62pal-1nob-1§
Misplaced hypodermal precursors, early embryo 
Misplaced hypodermal precursors, late embryo ± 
Enclosure defects, ventral cleft 
Enclosure defects, ventral hypodermis ± 
P→A lineage transformations 
Defective gene
Phenotypeunc-62pal-1nob-1§
Misplaced hypodermal precursors, early embryo 
Misplaced hypodermal precursors, late embryo ± 
Enclosure defects, ventral cleft 
Enclosure defects, ventral hypodermis ± 
P→A lineage transformations 
*

Based on results in this paper, previously published work(Edgar et al., 2001;Van Auken et al., 2000;Van Auken, 1998) and unpublished results.

Homozygous mutant progeny embryos of s472/+ heterozygous hermaphrodites or progeny embryos of t2012 homozygous hermaphrodites.

Homozygous mutant progeny embryos of pal-1(ct224)/+ heterozygous hermaphrodites; early pal-1 functions (Hunter and Kenyon, 1996) are maternally supplied (Edgar et al.,2001).

§

These strong lf alleles are deletions that probably eliminate function of both nob-1 and php-3, whose functional relationships are not yet clear (Van Auken et al., 2000).

These phenotypes are less severe than in unc-62(lf) andpal-1(lf) embryos (see text).

Are UNC-62 proteins required to mediate nuclear localization of CEH-20, as might be expected from findings in other organisms? It appears that this requirement may be partial rather than complete; unc-62-defective embryos show variable nuclear localization of a translational CEH-20::GFP reporter in CEH-20-expressing cells, with less localization in early than in late embryos [see supplemental Fig. S4],as if other mechanisms of CEH-20 nuclear import may exist.

Although the identity and interactions of all the players are not yet clear, our results do suggest that embryonically acting unc-62products are involved in controlling hypodermal specification, differentiation and morphogenesis. In wild-type embryos, dorsal and lateral hypodermal cell precursors form distinct populations that reside next to each other initially on the dorsal side of the embryo. We observed early s472, t2012 andct344 mutant embryos in which dorsal and lateral hypodermal precursors intermingled, and fewer than the normal number of cells expressing hypodermal markers formed subsequently. These abnormalities probably contribute to later morphological defects, as successful enclosure, elongation and morphogenesis require proper hypodermal cell arrangement and function(Priess and Hirsh, 1986;Williams-Masson et al., 1998;Williams-Masson et al., 1997). The abnormal mixing of hypodermal cell populations in these mutants may indicate involvement of unc-62 in maintaining distinct hypodermal cell identities, perhaps similar to the role of Drosophila hth in specifying distinct populations of cells that normally do not intermingle in the leg (Wu and Cohen, 1999)and eye-antenna (Pichaud and Casares,2000) imaginal discs.

There are several differences between C. elegans andDrosophila that may somehow be related to our finding that phenotypes caused by unc-62, ceh-20 and ceh-40 mutations are not as expected from observations on the homologous Drosophila genes. Whereas in Drosophila only one PBC-family and one MEIS-family gene have been reported, C. elegans has at least two PBC-family genes. Furthermore, the unc-62 gene generates four different transcripts by alternative splicing, while Drosophila hth transcripts are not alternatively spliced. Therefore, C. elegans, with several PBC-family and Meis-family proteins, resembles vertebrates more closely thanDrosophila with regard to the repertoire and possible interactions of these cofactors. A second possibly relevant difference between C. elegans and Drosophila could be that C. elegansrequires only anterior-group and posterior-group Hox gene functions for embryonic survival (Brunschwig et al.,1999; Van Auken et al.,2000). As PBC- and Meis-family proteins can act as cofactors for Hox proteins, the fact that medial-group Hox proteins are essential for completion of embryogenesis in Drosophila but not C. elegansmight account for different phenotypes resulting from loss of co-factor functions.

Supplemental data available on-line

Acknowledgements

We are grateful to D. Baillie, R. Schnabel and A. Chisholm forunc-62 alleles; to M. Stern for ceh-20 alleles and unpublished results; to H. Kagoshima and T. Burglin for theceh-20::GFP construct; to Y. Kohara for cDNAs; to Yo Suzuki and Dan Escovitz for assistance with several experiments; and to J. Yoder for the wild-type image in Fig. 6A. We thank M. Han, in whose laboratory the ku234 allele was isolated, for support and encouragement; Andrew Taft and Chris Link for assistance with the QPCR experiments; and Lucie Yang and Cynthia Kenyon for pointing out to us the molecular lesion in ku234. We also thank the Genome Sequencing Center, Washington University, St Louis for communication of DNA sequence data prior to publication. Some strains were obtained from the Caenorhabditis Genetics Center, which is supported by the NIH Division of Research Resources. This research was supported by grants to W. B. W. (HD-14958) and M. S.(GM-58540) from NIH.

References

Abu-Shaar, M., Ryoo, H. D. and Mann, R. S.(
1999
). Control of the nuclear localization of Extradenticle by competing nuclear import and export signals.
Genes Dev.
13
,
935
-945.
Affolter, M., Marty, T. and Vigano, M. A.(
1999
). Balancing import and export in development.
Genes Dev.
13
,
913
-915.
Babu, P. (
1974
). Biochemical genetics ofCaenorhabditis elegans.
Mol. Gen. Genet.
135
,
39
-44.
Beanan, M. J. and Strome, S. (
1992
). Characterization of a germ-line proliferation mutation in C. elegans.
Development
116
,
755
-766.
Berthelsen, J., Kilstrup-Nielsen, C., Blasi, F., Mavilio, F. and Zappavigna, V. (
1999
). The subcellular localization of PBX1 and EXD proteins depends on nuclear import and export signals and is modulated by association with PREP1 and HTH.
Genes Dev.
13
,
946
-953.
Brenner, S. (
1974
). The genetics ofCaenorhabditis elegans.
Genetics
77
,
71
-94.
Brunschwig, K., Wittmann, C., Schnabel, R., Burglin, T. R.,Tobler, H. and Mueller, F. (
1999
). Anterior organization of the Caenorhabditis elegans embryo by the labial-like Hox geneceh-13.
Development
126
,
1537
-1546.
Burglin, T. R. (
1997
). Analysis of TALE superclass homeobox genes (MEIS, PBC, KNOX, Iroquois, TGIF) reveals a novel domain conserved between plants and animals.
Nucleic Acids Res.
25
,
4173
-4180.
Chan, S., Jaffe, L., Capovilla, M., Botas, J. and Mann, R.(
1994
). The DNA binding specificity of ultrabithorax is modulated by cooperative interactions with extradenticle, another homeoprotein.
Cell
78
,
603
-615.
Chisholm, A. (
1991
). Control of cell fate in the tail region of C. elegans by the gene egl-5.
Development
111
,
921
-932.
Clark, S., Chisholm, A. and Horvitz, H. (
1993
). Control of cell fates in the central body region of C. elegans by the homeobox gene lin-39.
Cell
74
,
43
-55.
Edgar, L. G., Carr, S., Wang, H. and Wood, W. B.(
2001
). Zygotic expression of the caudal homologpal-1 is required for posterior patterning in Caenorhabditis elegans embryogenesis.
Dev. Biol.
229
,
71
-88.
Ekker, S. C., Jackson, D. G., von Kessler, D. P., Sun, B. I.,Young, K. E. and Beachy, P. A. (
1994
). The degree of variation in DNA sequence recognition among four Drosophila homeotic proteins.
EMBO J.
13
,
3551
-3560.
Giulietti, A., Overbergh, L., Valckx, D., Decallonne, B.,Bouillon, R. and Mathieu, C. (
2001
). An overview of real-time quantitative PCR: applications to quantify cytokine gene expression.
Methods
25
,
386
-401.
Hoey, T. and Levine, M. (
1988
). Divergent homeo box proteins recognize similar DNA sequences in Drosophila.
Nature
332
,
858
-861.
Kimble, J., Edgar, L. and Hirsh, D. (
1984
). Specification of male development in Caenorhabditis elegans: thefem genes.
Dev. Biol.
105
,
234
-239.
Kurant, E., Eytan, D. and Salzberg, A. (
2001
). Mutational analysis of the Drosophila homothorax gene.
Genetics
157
,
689
-698.
Labouesse, M., Hartwieg, E. and Horvitz, H. R.(
1996
). The Caenorhabditis elegans LIN-26 protein is required to specify and/or maintain all non-neuronal ectodermal cell fates.
Development
122
,
2579
-2588.
Lipardi, C., Wei, Q. and Paterson, B. M.(
2001
). RNAi as random degradative PCR. siRNA primers convert mRNA into dsRNAs that are degraded to generate new siRNAs.
Cell
107
,
297
-307.
Liu, J. and Fire, A. (
2000
). Overlapping roles of two Hox genes and the exd ortholog ceh-20 in diversification of the C. elegans postembryonic mesoderm.
Development
127
,
5179
-5190.
Livak, K. J. and Schmittgen, T. D. (
2001
). Analysis of relative gene expression data using real-time quantitative PCR and the 2(—Delta Delta C(T)) method.
Methods
25
,
402
-408.
Manak, J. and Scott, M. (
1994
). A class act:conservation of homeodomain protein functions.
Development
Suppl.
,
61
-71.
Mann, R. S. (
1995
). The specificity of homeotic gene function.
BioEssays
17
,
855
-863.
Mann, R. S. and Affolter, M. (
1998
).Hox proteins meet more partners.
Curr. Opin. Genet. Dev.
8
,
423
-429.
Mann, R. S. and Chan, S. K. (
1996
). Extra specificity from extradenticle: the partnership between HOX and PBX/EXD homeodomain proteins.
Trends Genet.
12
,
258
-262.
Montgomery, M. K., Xu, S. and Fire, A. (
1998
). RNA as a target of double-stranded RNA-mediated genetic interference inCaenorhabditis elegans.
Proc. Natl. Acad. Sci. USA
95
,
15502
-15507.
Pai, C. Y., Kuo, T. S., Jaw, T. J., Kurant, E., Chen, C. T.,Bessarab, D. A., Salzberg, A. and Sun, Y. H. (
1998
). The Homothorax homeoprotein activates the nuclear localization of another homeoprotein, extradenticle, and suppresses eye development in Drosophila.
Genes Dev.
12
,
435
-446.
Pichaud, F. and Casares, F. (
2000
).homothorax and iroquois-C genes are required for the establishment of territories within the developing eye disc.
Mech. Dev.
96
,
15
-25.
Priess, J. R. and Hirsh, D. I. (
1986
).Caenorhabditis elegans morphogenesis: The role of the cytoskeleton in elongation of the embryo.
Dev. Biol.
117
,
156
-173.
Pulak, R. and Anderson, P. (
1993
). mRNA surveillance by the Caenorhabditis elegans smg genes.
Genes Dev.
7
,
1885
-1897.
Rieckhof, G. E., Casares, F., Ryoo, H. D., Abu-Shaar, M. and Mann, R. S. (
1997
). Nuclear translocation of Extradenticle requires Homothorax, which encodes an Extradenticle-related homeodomain protein.
Cell
91
,
171
-183.
Rosenbluth, R., Rogalski, T., Johnson, R., Addison, L. and Baillie, D. (
1988
). Genomic organization inCaenorhabditis elegans: deficiency mapping on linkage group V(left).
Genet. Res.
52
,
105
-118.
Ryoo, H. D., Marty, T., Casares, F., Affolter, M. and Mann, R. S. (
1999
). Regulation of Hox target genes by a DNA bound Homothorax/Hox/Extradenticle complex.
Development
126
,
5137
-5148.
Schauer, I. and Wood, W. (
1990
). Early C. elegans embryos are transcriptionally active.
Development
110
,
1303
-1317.
Shen, W. F., Montgomery, J. C., Rozenfeld, S., Moskow, J. J.,Lawrence, H. J., Buchberg, A. M. and Largman, C. (
1997a
). AbdB-like Hox proteins stabilize DNA binding by the Meis 1 homeodomain proteins.
Mol. Cell. Biol.
17
,
6448
-6458.
Shen, W. F., Rozenfeld, S., Lawrence, H. J. and Largman, C.(
1997b
). The Abd-B-like Hox homeodomain proteins can be subdivided by the ability to form complexes with Pbx1a on a novel DNA target.
J. Biol. Chem.
272
,
8198
-8206.
Siddiqui, S. (
1990
). Mutations affecting axonal growth and guidance of motor neurons and mechanosensory neurons in the nematode Caenorhabditis elegans.
Neurosci. Res.
13
,
171
-190.
Streit, A., Li, W., Robertson, B., Schein, J., Kamal, I. H.,Marra, M. and Wood, W. B. (
1999
). Homologs of theCaenorhabditis elegans masculinizing gene her-1 in C. briggsae and the filarial parasite Brugia malayi.
Genetics
152
,
1573
-1584.
Sulston, J. and Hodgkin, J. (
1988
). Methods. In
The Nematode Caenorhabditis elegans
(ed. W. B. Wood),pp.
81
-122. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press.
Tabara, H., Grishok, A. and Mello, C. C.(
1998
). RNAi in C. elegans: soaking in the genome sequence.
Science
282
,
430
-431.
Takacs-Vellai, K., Vellai, T., Vigano, A., Affolter, M. and Mueller, F. (
2002
). A functional analysis of theextradenticle ortholog ceh-40.
Abstracts, European C. elegans meeting
, Paestum, Italy, May 18-21.,
197
.
Thomas, C., DeVries, P., Hardin, J. and White, J.(
1996
). Four-dimensional imaging: computer visualization of 3D movements in living specimens.
Science
273
,
603
-607.
Thomas, J., Stern, M. and Horvitz, H. (
1990
). Cell interactions coordinate the development of the C. elegansegg-laying system.
Cell
62
,
1041
-1052.
Trent, C., Tsung, N. and Horvitz, H. R. (
1983
). Egg-laying defective mutants of the nematode Caenorhabditis elegans.
Genetics
104
,
619
-647.
Van Auken, K., Weaver, D. C., Edgar, L. G. and Wood, W. B.(
2000
). C. elegans embryonic axial patterning requires two recently discovered posterior-group Hox genes.
Proc. Natl. Acad. Sci. USA
97
,
4499
-4503.
Van Auken, K. M. (
1998
).
Genetic and molecular analysis of nob-1, a gene required for posterior development in the Caenorhabditis elegans embryo
. PhD dissertation, University of Colorado, Boulder, CO.
van Dijk, M. and Murre, C. (
1994
). Extradenticle raises the Dna binding specificity of homeotic selector gene products.
Cell
78
,
617
-624.
Wang, B., Muller-Immergluck, M., Austin, J., Robinson, N.,Chisholm, A. and Kenyon, C. (
1993
). A homeotic gene cluster patterns the anteroposterior body axis of C. elegans.
Cell
74
,
29
-42.
Williams-Masson, E. M., Heid, P. J., Lavin, C. A. and Hardin,J. (
1998
). The cellular mechanism of epithelial rearrangement during morphogenesis of the Caenorhabditis elegans dorsal hypodermis.
Dev. Biol.
204
,
263
-276.
Williams-Masson, E. M., Malik, A. N. and Hardin, J.(
1997
). An actin-mediated two-step mechanism is required for ventral enclosure of the C. elegans hypodermis.
Development
124
,
2889
-2901.
Wrischnik, L. A. and Kenyon, C. J. (
1997
). The role of lin-22, a hairy/enhancer of split homolog, in patterning the peripheral nervous system of C. elegans.
Development
124
,
2875
-2888.
Wu, J. and Cohen, S. M. (
1999
). Proximodistal axis formation in the Drosophila leg: subdivision into proximal and distal domains by Homothorax and Distal-less.
Development
126
,
109
-117.

Supplementary information