The gene mab-21, which encodes a novel protein of 386 amino acids, is required for the choice of alternate cell fates by several cells in the C. elegans male tail. Three cells descended from the ray 6 precursor cell adopt fates of anterior homologs, and a fourth, lineally unrelated hypodermal cell is transformed into a neuroblast. The affected cells lie together in the lateral tail epidermis, suggesting that mab-21 acts as part of a short-range pattern-formation mechanism. Each of the changes in cell fate brought about by mab-21 mutants can be interpreted as a posterior-toanterior homeotic transformation. mab-21 mutant males and hermaphrodites have additional pleiotropic phenotypes affecting movement, body shape and fecundity, indicating that mab-21 has functions outside the tail region of males. We show that the three known alleles of mab-21 are hypomorphs of a new gene. Mosaic analysis revealed that mab-21 acts cell autonomously to specify the properties of the sensory ray, but non-autonomously in the hypodermal versus neuroblast cell fate choice. Presence of cell signalling in the choice of the neuroblast fate was confirmed by cell ablation experiments. Mutations in mab-21 were shown previously to be genetic modifiers of the effects of HOMC/Hox gene mutations on ray identity specification. The results presented here support the conclusion that mab-21 acts as part of a mechanism required for correct cell fate choice, possibly involving the function of HOM-C/Hox genes in several body regions.

During animal development, pattern-formation mechanisms guide the generation of variant forms of a variety of serially repeated structures, such as segments, rhombomeres, digits or sensory organs. One way in which these pattern-formation mechanisms exert their influence is by determining unique states of expression of HOM-C/Hox transcription factors within cells (McGinnis and Krumlauf, 1992). HOM-C/Hox transcription factors, in turn, are thought to regulate transcription of downstream genes that determine the individual properties of separate units (Andrew and Scott, 1992; Botas, 1993). In C. elegans, HOM-C/Hox genes determine variant forms of epidermal and neuronal cell sublineages that are serially repeated along the anteroposterior body axis (Sulston and Horvitz, 1977; Wang et al., 1993). As one example of this process, we have studied development of a set of nine bilateral pairs of peripheral sense organs known as rays present in the posterior region of males. Each ray develops from an identical ray cell sublineage, yet each ray develops at a different epidermal site, can have a distinct morphology and pattern of neurotransmitter expression, and can mediate distinct behavioral responses during mating (Sulston and Horvitz, 1977; Sulston et al., 1980; Loer and Kenyon, 1993; Liu and Sternberg, 1995). In addition, the cells of each ray appear to express a unique combination of cell recognition functions necessary for their assembly as a separate organ. Mutations in several genes cause neighboring rays to fuse together during development, suggesting that they have lost their distinct identities and instead express common assembly functions (Baird et al., 1991).

We have shown that HOM-C/Hox genes play a role in specifying ray identities (Chow and Emmons, 1994). Several of the ray fusion genes that we identified are genetic modifiers of HOM-C/Hox effects on ray development (Chow and Emmons, 1994), suggesting that ray fusion genes act in a common pathway with HOM-C/Hox genes (e.g. Botas et al., 1982; Kusch and Edgar, 1986; Simon et al., 1991). In the present paper, we characterize further one of these HOM-C/Hox gene modifier loci.

Mutations in the gene mab-21 result in the specific transformation of the identity of one of the rays into that of a more anterior ray (Baird et al., 1991). Here we characterize this transformation in greater detail and document effects of mab-21 mutations on the cell fate choices and differentiation of four cells. The results are consistent with a model in which mab-21 acts as part of a localized pattern-formation mechanism that dictates the fates of cells in the male tail lateral hypodermis. It could as well have a wider function in the specification by HOM-C/Hox genes of specialized structures or cell identities elsewhere in the body.

General methods and mapping

Nematodes were reared at 20°C following standard procedures (Brenner, 1974). Most strains carried the him-5(e1490) mutation, which results in a high frequency of spontaneous males in populations of selfing hermaphrodites (Hodgkin et al., 1979). Isolation after EMS mutagenesis of two alleles of mab-21 (bx41 and bx53) has been described previously (Baird et al., 1991); a third allele (sy155) was isolated from a mutant strain kindly provided by H. Chamberlin and P. Sternberg. mab-21 mutations were mapped to the left arm of linkage group III between unc-79 and pal-1 by a four factor cross in which the following recombination events were scored: unc-79 (12/18) mab-21 (3/18) pal-1 (3/18) dpy-17 (see Fig. 4A). mab-21 was placed between two restriction fragment length polymorphisms, MJ#NEC2 and MJ#NEC1, by crossing a mab-21 strain to strain EM375 of genotype unc-79 MJ#NEC2 MJ#NEC1 dpy-17 with the following recombination results: unc-79 (29/59) MJ#NEC2 (11/59) mab-21 (9/59) MJ#NEC1 (10/59) dpy-17 (Fig. 4B). EM375 was generated from strain MJ569 provided by J. Miwa.

Cell lineage analysis and cell ablation experiments

Cell lineages of mab-21(bx41) (3 sides) and mab-21(bx53) (5 sides) were followed in living animals by Nomarski microscopy following methods described by Sulston and Horvitz (1977). Lineages were followed from late L1 through the L3 molt and division of Rn.a cells. Division of Rn.a cells was taken as an indication of the expression of the ray sublineage. Cells were ablated with a laser microbeam following procedures described in Chow and Emmons (1994).

Electron microscopy

Young adult males were rinsed in M9 buffer and placed into buffered aldehyde fixative at room temperature. While in fix, the tails were quickly cut off with a scalpel blade. After 1-3 hours in aldehyde fix, samples were transferred through several buffer rinses and refixed in 1% osmium tetroxide. Samples were then dehydrated and embedded in Medcast resin (Ted Pella) and serially sectioned on a diamond knife (cf. Hall, 1995). Some tails were sectioned transverse to the body axis, others lengthwise, either sagittal or frontal, in order to obtain different perspectives of the curving rays. Electron micrographs were collected with a Philips CM10 at 2,000-12,000×. Several rays that showed especially good fixation were reconstructed using the Eutectics Electronics 3D-SSRS computer program, after manual tracing of every 3-5th section on a digitizing tablet. In mab-21 males, normal rays, an ectopic ray and a fused (4+6) ray could thus be compared from any perspective and used as models with which to compare sample sections from additional animals. In all, nine mab-21 adults and four wild-type adults were compared in serial thin sections. In particular, effort was made to compare the details of ray openings, ciliary tip specializations, the relative sizes of cellular elements along the length of the rays, and the number and type of cell processes invading the base of each ray. Not all features could be assessed in each specimen, depending upon section angle and varying quality of preservation.

Complementation of emb genes

The following embryonic lethal mutations, all temperature sensitive and closely linked to mab-21, were tested for complementation with mab-21: emb-1(hc57), emb-2(hc58), emb-5(hc61), emb-7(hc66), emb-8(hc69), emb-13(g6) and emb-32(g58). Crosses were set up at 16°C between mab-21(bx53);him-5(e1490) males and emb-n hermaphrodites. After mating took place for 2 days, the parents were removed, and plates with mostly eggs and a few L1 larvae were transferred to 25°C. mab-21/emb-n male cross progeny were scored for ray fusion phenotype.

Indirect immunofluorescence staining of cell boundaries

Animals were fixed with 1% paraformaldehyde and permeabilized by reduction and oxidation of the cuticle (Finney and Ruvkun, 1990). Permeabilized animals were incubated at 37°C for 8 hours with a hundred-fold dilution of monoclonal antibody MH27 (provided by J. Preiss) and then for 12 hours with a hundred-fold dilution of secondary antibody (rhodamine isothiocyanate conjugate goat antimouse, Boehringer Mannheim Biochemical, Indianapolis IN). Stained animals were mounted in a solution containing 30 mM Tris pH7.5, 70% glycerol and 2% N-propylgalate. Observation and photography were performed with a Zeiss Axioplan microscope equipped for epifluorescence.

Visualization of hypodermal cell boundaries

To visualize the SET compartment in the male tail, late L4 larval males were mounted for Nomarski microscopy (Sulston and Hodgkin, 1988) in 0.25% sodium dodecyl-sulfate (Austin and Kenyon, 1994). Animals were immediately observed at 1000×. The boundary of hypodermal cells, including the SET, gradually became visible and then started to deteriorate. When the boundaries were clear, SET nuclei were counted and photographed.

Transformation rescue of mab-21

Cosmids for transformation rescue of mab-21 were obtained from R. Shownkeen and A. Coulson (MRC Laboratory of Molecular Biology, Cambridge, UK). Microinjection of cloned DNA into hermaphrodite gonads was performed following the procedure of Mello et al. (1991). A number of cosmids including those listed in Fig. 4B were individually injected at a concentration of 15 μg/ml together with the dominant rol-6(su1006) marker carried on plasmid pRF4 (Kramer et al., 1990) at a concentration of 20 μg/ml. mab-21(bx53);him-5(e1490) (EM128) was the host for microinjection. Rol males appearing among the progeny of injected hermaphrodites were scored for ray 6 to 4 transformation phenotype. A minimal rescuing genomic BamHI-SacI fragment was identified by deletion analysis (EM#227) (Fig. 4B) and its sequence determined (Sequenase Kit, US Biochemicals).

To confirm the identification of the mab-21 locus, a frame shift mutation was introduced into the open reading frame present in EM#219 (Fig. 4B). A unique EcoRI site in the 3′ end of the open reading frame was filled in with Klenow polymerase and religated. This resulted in generation of a premature translation termination codon at position 311 and, consequently, a truncated MAB-21 conceptual protein lacking 79 amino acids from the carboxyl terminus. Transformation with DNA carrying this frame shift mutation failed to result in rescue of mab-21(bx53) (0/54 Rol males showed a wild-type ray 6 phenotype).

Isolation of cDNA clones and heat-shock-induced expression

A C. elegans λ cDNA library (Palazzolo et al., 1990) was screened with EM#219 (Fig. 4B). Plasmid derivatives of the positive phage clones were obtained following the procedure described by Palazzolo et al. (1990). Inserts were sequenced with Sequenase Kit (US Biochemicals). DNA sequences were analyzed with GCG sequence analysis software (Sequence Analysis Software Package by Genetics Computer Group, Inc., [Devereux et al., 1984]). The GenBank Accession Number for the mab-21 cDNA sequence is U19861

cDNA inserts were released by digestion with ApaI and SacI and subcloned into pPD49.78 (Fire et al., 1990), which allows their expression from the heat-shock promoter hsp16.2 (generating respectively plasmids EM#230, EM#231 and EM#220). After cotransformation with pRF4, stable integrated lines were isolated by γ ray irradiation (95 rads per minute for 40 minutes). For heat shock of synchronized populations, eggs were isolated with sodium hypochlorite and allowed to hatch into buffer (Sulston and Hodgkin, 1988). Synchronized, arrested L1 larvae were transferred after 24 hours to 50 mm plates seeded with bacteria and allowed to develop at 20°C. At various times during development, plates were shifted to 32°C for 3 hours and returned to 20°C. Adult Rol males were scored for ray fusion phenotype. For determining the effective heat-shock interval with individual animals, single animals were randomly picked and staged by observation with Nomarski microscopy. They were then heat shocked for 3 hours at 32°C, allowed to recover and develop at 20°C, and their adult phenotype recorded (Fig. 5).

Mosaic analysis

Mosaic analysis (Herman, 1984, 1989) for determination of the focus of mab-21 gene activity was carried out in strain EM289, genotype mab-21(bx53)ncl-1(e1865)unc-36(e251)III;him-5(e1490)V; sDp3(III;f). The free duplication sDp3(III;f) (Rosenbluth et al., 1985) carries wild-type copies of the LGIII genes, and hence EM289 is generally non-Mab non-Unc and non-Ncl in phenotype. Genetic mosaics arise by spontaneous loss of the free duplication. ncl-1(e1865) results in enlarged nucleoli; its activity is cellautonomous (Herman, 1989). The unc-36(e251) mutation results in very slow movement, probably due to loss of the gene in motor neurons; its focus of activity is among descendants of AB or AB.p (Kenyon, 1986; Herman, 1989). Mosaic males were identified as Unc animals with wild-type rays, or as non-Unc or semi-Unc animals with Mab rays. Because displacement of ray cell bodies during L4 morphogenesis of the male tail prevents identification of ray cells in the adult, the most likely point of duplication loss was determined by examining the nuclei of lineally related cells (Fig. 6). The Ncl phenotype of PLM(L/R), V5.pa postdeirid neurons and Q-derived neurons were scored to infer duplication loss in the AB.p branch; the Ncl phenotype in ALM(L/R) and BDU(L/R) were scored to infer loss in the AB.a branch. Loss of the duplication in the P1 lineage was scored by the Ncl phenotype of body muscle cells, pharyngeal neurons and coelomocytes. In four putative mosaic animals, no Ncl nucleoli could be found in the cells examined. The mab-21 phenotype observed on one side in these animals could be accounted for by loss of the duplication in a subset of cells within the V6 lineage (in two cases within AB.alpppp or one of its descendants, in two cases within AB.arpppp or one of its descendants); these are not shown in Fig. 6.

mab-21 is required for the choice of alternate fates by several cells in the male tail

Mutations in the gene mab-21 were identified because of their effects on differentiation of a single pair of sensory rays in the C. elegans adult male tail (Baird et al., 1991). We found that mab-21 mutations affected the differentiation of two cells in ray 6, as well as the fusion properties of the hypodermal cell generated by the ray 6 sublineage. mab-21 mutations also transform a nearby hypodermal cell into a neuroblast cell that expresses the ray sublineage and generates a ray.

The C. elegans male has several specialized posterior structures and organs necessary for copulation with the hermaphrodite. Among these are a set of nine bilaterally symmetrical pairs of sensory rays, which project outward from the body within an acellular fan (Fig. 1A). The ultrastructure of the rays has been described by Sulston et al. (1980) and is illustrated in Fig. 2. Each ray consists of a cone or cylinder of hypodermis containing the dendritic processes of two ultrastructurally distinguishable neurons (RnA, RnB, n=1-9), together with the process of a support cell (called the structural cell, Rnst) (Fig. 2J,N). The support cell is joined to the hypodermis at the tip of the ray and surrounds and holds the dendritic endings of the neurons, one of which, RnB, generally faces an opening to the exterior (Fig. 2A,D). In wild-type males, the rays are of similar morphology with one exception: the sixth ray counting from anterior to posterior, ray 6, is fatter and more conical in shape than the other rays (Fig. 1A). Ray 6 also fails to open to the exterior or is open only through a thin channel and has a B neuron that differs slightly in ultrastructure from the B neurons of the other rays. (Sulston et al., 1980) (Fig. 2C).

Fig. 1.

Morphology and development of wild-type and mab-21 male tails. (A) Wild-type C. elegans male tail, ventral view, showing fan and rays. Arrow, ray 4; arrowhead, ray 6. Nomarski photomicrograph, scale bar, 10 μm. (B) mab-21 mutant male tail. Arrow indicates large ray consisting of a fusion of rays 4 and 6; E indicates the ‘ectopic’ tenth ray. (C) Arrangement of seam cells in the lateral epidermis of the male tail, late L3 larval stage. Rn cells (n=1-9) are ray precursor cells; T.apapa is a hypodermal cell. Most of the remainder of the surface is covered by a hypodermal syncytium. The lineal relationships of cells affected by mab-21 mutations are indicated. (D) Arrangement of seam cells approximately 5 hours later than the arrangement shown in C. At this time, all the Rn cells in wild type express the ray sublineage, shown only for R6 and R8. T.apapa fuses with the surrounding hypodermal syncytium. A, RnA; B, RnB; st, Rnst; hyp, hypodermal cell; X, programmed cell death. For description of these cells, see text.

Fig. 1.

Morphology and development of wild-type and mab-21 male tails. (A) Wild-type C. elegans male tail, ventral view, showing fan and rays. Arrow, ray 4; arrowhead, ray 6. Nomarski photomicrograph, scale bar, 10 μm. (B) mab-21 mutant male tail. Arrow indicates large ray consisting of a fusion of rays 4 and 6; E indicates the ‘ectopic’ tenth ray. (C) Arrangement of seam cells in the lateral epidermis of the male tail, late L3 larval stage. Rn cells (n=1-9) are ray precursor cells; T.apapa is a hypodermal cell. Most of the remainder of the surface is covered by a hypodermal syncytium. The lineal relationships of cells affected by mab-21 mutations are indicated. (D) Arrangement of seam cells approximately 5 hours later than the arrangement shown in C. At this time, all the Rn cells in wild type express the ray sublineage, shown only for R6 and R8. T.apapa fuses with the surrounding hypodermal syncytium. A, RnA; B, RnB; st, Rnst; hyp, hypodermal cell; X, programmed cell death. For description of these cells, see text.

Fig. 2.

Ultrastructure of wild-type and fused rays. (A,B) Tip of ray 4 in wild type. The dendritic ending of R4B narrows and is exposed to the exterior (thin arrow). It is surrounded by a surface density at the point of narrowing (thick arrow). (C) Cross section near the tip of ray 6 in wild type. This and adjacent sections through the tip reveal no surface density surrounding the dendritic ending of R6B. (D) Diagram of the tip of a typical ray in wild type other than ray 6, viewed from the dorsal or ventral side (perpendicular to the fan). The density surrounding RnB near the tip is indicated by an arrow, the edge of the fan by an arrowhead. A channel within the structural cell encloses the distal portion of both neuronal processes. Ray 6 is similar but lacks the open channel (or has only a very narrow channel through the structural cell not containing the ending of a neuron). The plane of section in A and B is indicated. (E) Diagram of the tip of a fused 4–6 ray in mab-21. Generally, three dendritic processes are present in two channels (n=13/14), RnA being missing from one channel. In one animal, a fourth dendritic process was present in a third channel. The planes of sections shown in F-I are indicated. (F-I) Sections through a fused 4-6 ray in mab-21. Two channels are present in a single structural cell, one with two neuronal processes and one with one. A process with the ultrastructure of a B-type neuron, each with a density similar to that found in ray 4 in wild type (arrows in G and H), extends into each of two openings. The levels of these sections are shown in E. (J) Schematic diagrams of complete rays, a fused 4-6 ray (top) and a normal ray (bottom). In the one fused ray reconstructed, the A-type neuron partially entered the ray as shown (thin arrow). A nucleus typically present within the base of the fused ray is indicated (arrowhead). The levels of the sections shown in K and M are indicated. (K,L) Representative cross section of a fused 4–6 ray in mab-21. The processes of 4 neurons, two B type, one A-type and one presumed A-type that fails to reach the tip, are present, as well as the process of a single structural cell. The bulk of the ray consists of the hypodermal syncytium, hyp7. The level of the section is indicated in J. (M,N) Representative cross section of the ‘ectopic’ ray generated by T.apapa in mab-21, showing that it has normal structure, typical of rays 1–5, 7–9 in wild type. The level of this section is indicated in J. Scale bars: in C, applies to A-C; in F applies to F-I; in M, applies to K-N. Definitions: hyp, hypodermis; str, structural cell; ext, exterior; fan, cuticular fan; RnA” and RnB”, second pair of neurons in fused ray.

Fig. 2.

Ultrastructure of wild-type and fused rays. (A,B) Tip of ray 4 in wild type. The dendritic ending of R4B narrows and is exposed to the exterior (thin arrow). It is surrounded by a surface density at the point of narrowing (thick arrow). (C) Cross section near the tip of ray 6 in wild type. This and adjacent sections through the tip reveal no surface density surrounding the dendritic ending of R6B. (D) Diagram of the tip of a typical ray in wild type other than ray 6, viewed from the dorsal or ventral side (perpendicular to the fan). The density surrounding RnB near the tip is indicated by an arrow, the edge of the fan by an arrowhead. A channel within the structural cell encloses the distal portion of both neuronal processes. Ray 6 is similar but lacks the open channel (or has only a very narrow channel through the structural cell not containing the ending of a neuron). The plane of section in A and B is indicated. (E) Diagram of the tip of a fused 4–6 ray in mab-21. Generally, three dendritic processes are present in two channels (n=13/14), RnA being missing from one channel. In one animal, a fourth dendritic process was present in a third channel. The planes of sections shown in F-I are indicated. (F-I) Sections through a fused 4-6 ray in mab-21. Two channels are present in a single structural cell, one with two neuronal processes and one with one. A process with the ultrastructure of a B-type neuron, each with a density similar to that found in ray 4 in wild type (arrows in G and H), extends into each of two openings. The levels of these sections are shown in E. (J) Schematic diagrams of complete rays, a fused 4-6 ray (top) and a normal ray (bottom). In the one fused ray reconstructed, the A-type neuron partially entered the ray as shown (thin arrow). A nucleus typically present within the base of the fused ray is indicated (arrowhead). The levels of the sections shown in K and M are indicated. (K,L) Representative cross section of a fused 4–6 ray in mab-21. The processes of 4 neurons, two B type, one A-type and one presumed A-type that fails to reach the tip, are present, as well as the process of a single structural cell. The bulk of the ray consists of the hypodermal syncytium, hyp7. The level of the section is indicated in J. (M,N) Representative cross section of the ‘ectopic’ ray generated by T.apapa in mab-21, showing that it has normal structure, typical of rays 1–5, 7–9 in wild type. The level of this section is indicated in J. Scale bars: in C, applies to A-C; in F applies to F-I; in M, applies to K-N. Definitions: hyp, hypodermis; str, structural cell; ext, exterior; fan, cuticular fan; RnA” and RnB”, second pair of neurons in fused ray.

In mab-21 mutants, the most obvious phenotype is that the distinctive, conical ray 6 is absent and a large ray of uniform diameter replaces ray 4 (Fig. 1B). The large new ray consists of a fusion of rays 6 and 4, and contains the cells normally found in these two rays, as demonstrated by several lines of evidence. First, in mab-21 mutants, the cell lineages leading to rays 4 and 6 are unaffected. In wild type, the three cells of each ray plus one hypodermal cell are generated as products of the ray sublineage. The ray sublineage is expressed by ray precursor cells (Rn cells) (Fig. 1C,D) (Sulston and Horvitz, 1977; Sulston et al., 1980). In mab-21 mutants, both R4 and R6 express the ray sublineage at the normal time (8/8 sides lineaged). Secondly, the large ray at the position of ray 4 extends from two adjacent papillae (not shown) and has two ray tips (Fig. 2F). Ray papillae and tips are structures formed by the ray structural cell and have a distinctive ‘ring-and-dot’ morphology visible by Nomarski microscopy (Sulston et al., 1980). The presence of two ray papillae during L4 and two tips in the large ray suggests that the large ray contains two structural cells. These cells are apparently fused, as no cell boundaries separating them are visible in electron micrographs of the fused ray (Fig. 2F-I,K). Thirdly, electron microscopy reveals that the large ray contains the processes of three or four neurons (Fig. 2F-I,K). Two of the processes terminate respectively at each of two openings and have the ultrastructure of B-type neurons, while two terminate within the ray and have the ultrastructure of A-type neurons.

Fusion of ray 6 to ray 4 in mab-21 mutants appears to be the result of altered properties of ray 6 and, in particular, of the ray 6 structural cell, R6st. Mutational changes affecting ray 6 alone can occasionally be seen when fusion of ray 6 with ray 4 does not occur. In 5% (n>600) of mab-21 mutant animals, rays 4 and 6 do not fuse. In these animals, ray 4 appears to be unaffected, while ray 6 lies between rays 4 and 5 anterior of its normal position. In such animals, ray 6 does not have a conical morphology, but instead has a uniformly thin morphology similar to the other rays. Similarly, after ablation of R6 in mab-21, no fused ray forms and a normal ray 4 is present, whereas after ablation of R4, ray 6 forms a thin ray between rays 3 and 5 (Table 3). In wild type, the unique conical morphology of ray 6, as well as its position between rays 5 and 7, is determined by R6st, as shown by ablation of the neurons (Zhang and Emmons, 1995). Altered ray 6 position in mab-21 is likely to result from altered interactions between R6st and surrounding hypodermal cells (Baird et al., 1991). Fusion of ray 4 and ray 6 in mab-21 mutants could be a consequence of misplacement of ray 6 to a position adjacent to ray 4, or could result because cell recognition functions expressed by cells of ray 6, ray 4, or both, are altered.

In addition to affecting properties of R6st, mab-21 mutations affect the ultrastructure of R6B. In wild type, the B neurons of most rays are characterized by having a thin tip that is exposed to the exterior and by the presence of an extracellular dense matrix material where the tip narrows (Fig. 2A,B). In ray 6 in wild type, in which there is either no opening or only a narrow channel opening to the exterior (data not shown), the B neuron lacks this dense material (Fig. 2C). In mab-21 mutants, fused rays have two openings, each containing the tip of a B-type neuron. Both of these neurons are surrounded by the dense material normally characteristic of all the rays except ray 6 (Fig. 2G,H). Thus in mab-21 mutants, the ultrastructure of the ending of R6B has been transformed to that characteristic of the B neurons of the other rays. A second unique ultrastructural property of R6B in wild type, lack of dilated cisternae in the cell body (Sulston et al., 1980), was not examined.

A third cell affected in mab-21 mutants is the hypodermal cell generated by the ray 6 sublineage (R6.p). At the end of the L4 larval stage, R6.p normally fuses with the large hypodermal syncytium, hyp7, covering most of the body (Sulston et al., 1980). In mab-21 mutants, R6.p instead fuses with R1.p-R5.p (n=15/15), thus becoming part of the tail seam (SET) (Fig. 3).

Fig. 3.

In mab-21, R6.p fuses with the tail seam (SET). (A-D) Photomicrographs and diagrams of cell boundaries in the left lateral tail hypodermis, visualized by indirect immunofluorescence staining with MH27 antibody. In mab-21, the cell boundaries delineating R6.p, which normally contact ray 5, have disappeared, and ray 5 lies within the SET. An additional Rn.p cell (T.apapap) and an ectopic ray (E) are present. Ph, phasmid. (E,F) Visualization of cell boundaries in the left lateral tail hypodermis by SDS treatment. In wild type, there are 5 nuclei within the SET and R6.p can be seen as a separate cell. In mab-21, there are 6 nuclei in the SET. Bar, 10 μm.

Fig. 3.

In mab-21, R6.p fuses with the tail seam (SET). (A-D) Photomicrographs and diagrams of cell boundaries in the left lateral tail hypodermis, visualized by indirect immunofluorescence staining with MH27 antibody. In mab-21, the cell boundaries delineating R6.p, which normally contact ray 5, have disappeared, and ray 5 lies within the SET. An additional Rn.p cell (T.apapap) and an ectopic ray (E) are present. Ph, phasmid. (E,F) Visualization of cell boundaries in the left lateral tail hypodermis by SDS treatment. In wild type, there are 5 nuclei within the SET and R6.p can be seen as a separate cell. In mab-21, there are 6 nuclei in the SET. Bar, 10 μm.

Finally, in mab-21 mutants, a hypodermal cell lying adjacent to ray 6 is transformed into a ray neuroblast. In wild type, the rays are generated by nine pairs of ray precursor cells (Fig. 1C). Ray precursor cells are descended from three embryonic hypodermal blast cells denoted V5 (ray 1), V6 (rays 2–6), and T (rays 7–9). In mab-21 mutants, an additional descendant of T, T.apapa, the anterior sister of R7, is often (45%, n>600) transformed from a hypodermal cell into a ray precursor cell and expresses the ray sublineage. Thus, mab-21 mutant animals can have 10 instead of 9 rays (Fig. 1B). The T.apapa-derived ray will sometimes be referred to as the ‘ectopic’ray. The ectopic ray is shown in cross section in Fig. 2M,N; it has the same general organization as other normal thin rays.

Altogether, the mab-21 male tail phenotype can be accounted for by the effects on the fates of the four cells, R6st, R6B, R6.p and T.apapa. These affected cells are of diverse types and therefore mab-21 does not appear to be required for expression of any particular cell fate. In each case, the affected cells choose alternate cell fates normally assumed by neighboring cells and appear to execute those alternate fates correctly. The cells affected by mab-21 mutations have in common that they lie adjacent to one another in a restricted region of the lateral hypodermis of the male tail, and differentiate during the late L3 and L4 larval stages. Therefore, it appears most likely that mab-21 is required for a localized pattern-formation mechanism that causes these cells to choose their wild-type fates.

All of the changes affecting cells of ray 6, together with the change in fusion partner of R6.p, can be parsimoniously interpreted as a change in the identity of the ray 6 precursor cell R6 to that of R4. T.apapa, like R6, is the anterior sister of a ray precursor cell (R7 in the case of T.apapa, R5 in the case of R6, Fig. 1C). R6 and T.apapa lie adjacent to each other ventral of the seam and, in mab-21 mutants, each assumes a new fate. mab-21 thus appears to act as one component of a patternformation mechanism that assigns fates to at least two seam cells during the late L3 larval stage.

Additional mab-21 mutant phenotypes suggest mab-21 has functions outside the tail region of males

In addition to having ray defects in the male tail, mab-21 males and hermaphrodites are somewhat short and fat, are slightly uncoordinated and have decreased fecundity. These pleiotropic defects, together with evidence described below that existing alleles are hypomorphic, imply that the normal action of mab-21 is not restricted to the tail hypodermis of males. The body length of mab-21 mutant adult hermaphrodites (856 μm±55 μm [n=30]), was intermediate between that of wild type (1024 μm±72 μm [n=30]) and dpy-17 (524 μm±56 μm [n=30]). The latter have short and fat bodies typical of a large number of dpy (dumpy) and sma (small) mutants. The shorter body length of mab-21 animals was not due to a growth rate defect, nor was the timing of developmental stages abnormal (data not shown). As the hypodermis and overlying cuticle play a key role in determining body shape in C. elegans (Priess and Hirsh, 1986; Levy et al., 1993) these results suggested that mab-21 might play a role in differentiation of the hypodermis. mab-21 animals, particularly males, have an uncoordinated movement in backward locomotion. Upon being touched on the head, mab-21 males made a ventral bend at the tail during backward movement that was more severe than that of wildtype males, resulting in a deep curving movement. A similar though less pronounced bend could be discerned in hermaphrodites. This backward Unc phenotype suggests a defect in the nervous system or possibly the posterior muscles in mab-21.

mab-21 hermaphrodites had a reduced brood size due to a decreased number of laid eggs, suggesting the presence of a gonadal defect (average number of self progeny at 20°C was 126±18 for mab-21(bx53);him-5(e1490), n=10, average number for him-5(e1490) was 221±18, n=10; results were similar with the other two mab-21 alleles). The eggs laid by mab-21 hermaphrodites failed to hatch in numbers (several percent) significantly higher than mab-21(+). Dead embryos were arrested from morphogenesis stage up to the end of embryogenesis, and had an apparently disrupted hypodermis. A few percent of hatched animals had a deformed Vab (variable abnormal) phenotype with deformities mostly in the region of the pharynx. mab-21 males sired fewer progeny in mating tests (data not shown), but it was difficult to determine whether this was due to a gonad defect or was because the uncoordinated movement described above disrupted mating.

mab-21 may have an essential embryonic function

In order to gain information about the possible null phenotype of mab-21, we determined the phenotype that resulted when alleles of mab-21 were placed over a deficiency. We attempted to construct the heterozygote carrying mab-21(bx53) over the deficiency yDf10, but found that progeny of the expected phenotype were absent from the cross (Table 1). In corresponding crosses lacking the mab-21 mutation, yDf10 heterozygotes were present in the expected proportion. We conclude that mab-21/yDf10 is a lethal combination and that the null phenotype of mab-21 is likely to be lethality prior to hatching. An alternative possibility that we cannot rule out is that another gene uncovered by the deficiency is haploinsufficient in the presence of a low amount of mab-21 gene product. Since the phenotype of mab-21(bx53) appears to be more severe over a deficiency and because the mutant phenotype of the other two mab-21 alleles is similar to that of bx53, we conclude that all these mutations are most likely hypomorphs.

Table 1.

Result of crossing mab-21 to a deficiency

Result of crossing mab-21 to a deficiency
Result of crossing mab-21 to a deficiency

We tested alleles of known essential genes in the vicinity of mab-21 for complementation with mab-21 (Schierenberg et al., 1980; Cassada et al., 1981) (Materials and Methods). Alleles of emb-1, emb-2, emb-5, emb-7, emb-8, emb-13 and emb-32 all complemented the ray 6 defect of mab-21(bx53) males. Therefore none of these mutations is likely to be a strong allele of mab-21 and mab-21 appears to define a new essential gene.

In order to gain information about the stage of embryonic arrest in a mab-21 near-null background, we examined the dead embryos arising from a cross between a mab-21-carrying strain and a yDf10-carrying strain, and compared these to dead embryos that segregated from a yDf10/+ strain during self propagation. 47% (36/76) of dead eggs resulting from the cross had completed the proliferation stage of embryogenesis and were arrested during the morphogenesis stage, whereas only 2% (1/56) of the dead eggs segregating from the yDf10/+ heterozygote progressed this far. Therefore, we conclude that mab-21/yDf10 heterozygous embryos can complete cellular proliferation and that decreased level of mab-21 function causes arrest during morphogenesis.

mab-21 encodes a novel protein

We cloned the mab-21 gene by identifying cosmids (kindly supplied by the C. elegans sequencing consortium) that rescued the mab-21 male tail phenotype. Candidate cosmids for testing were identified by mapping mab-21 between two restriction fragment length polymorphisms, MJ#NEC2 and MJ#NEC1 (Fig. 4B, see Materials and Methods). A 5.6 kb rescuing genomic fragment (EM#219) was used to screen a cDNA library (Palazzolo et al., 1990), resulting in the recovery of 8 cDNA clones from 250,000 plaques screened. Seven of the cDNA clones were of 1.3 kb and differed only by having 5′ ends located at different positions within a 26 bp genomic region. The eighth cDNA was of 0.7 kb and was colinear with the 3′ portion of the other cDNAs. We showed that these cDNAs represented transcripts of the mab-21 locus in two ways. First, expression of all of these cDNAs (except the 0.7 kb cDNA) under the control of a heat-shock promoter resulted in rescue of the mab-21 mutant phenotype, as described more fully below. Second, introduction of a mutation into the coding region of EM#219 abolished the ability of this fragment to rescue a mab-21 mutation (Materials and Methods). Third, a genomic fragment (EM#228) covering all of the genomic region covered by EM#227 with the exception of the region encoding the 3′ end of the cDNAs failed to rescue a mab-21 mutation (Fig. 4B).

Fig. 4.

(A) Genetic map of linkage group III in the region surrounding mab-21. (B) Physical map in the region surrounding mab-21. From top to bottom: relationship of physical and genetic markers; the number of recombination events between mab-21 and the two polymorphisms MJ#NEC2 and MJ#NEC1; YAC and cosmid clones and rescuing activity, those shown in bold rescued mab-21(bx53) with the frequencies indicated; rescuing BamHI restriction fragment subcloned from cosmid C56A8; a BamHI-SacI subclone with rescuing activity and EcoRI-BamHI and PstI-BamHI subclones without rescuing activity; the structure of a mab-21 transcript deduced by comparing genomic and cDNA sequences (open boxes represent exons, circle and arrow represent 5′ and 3′ ends of the isolated cDNAs, respectively). (C) Sequence of mab-21 genomic DNA. The amino acid sequence encoded by the mab-21 open reading frame, defined by cDNA sequences, is shown below in single letter code. The positions of introns are indicated by arrowheads; intronic sequences are not shown. Three mab-21 cDNA species begin with the underlined nucleotides indicated by the letters A, B and C. The putative polyadenylation signal is underlined and the polyadenylation site is indicated by an arrow. The EcoRI site used to generate a frame shift mutation for confirming the identification of mab-21 is labelled.

Fig. 4.

(A) Genetic map of linkage group III in the region surrounding mab-21. (B) Physical map in the region surrounding mab-21. From top to bottom: relationship of physical and genetic markers; the number of recombination events between mab-21 and the two polymorphisms MJ#NEC2 and MJ#NEC1; YAC and cosmid clones and rescuing activity, those shown in bold rescued mab-21(bx53) with the frequencies indicated; rescuing BamHI restriction fragment subcloned from cosmid C56A8; a BamHI-SacI subclone with rescuing activity and EcoRI-BamHI and PstI-BamHI subclones without rescuing activity; the structure of a mab-21 transcript deduced by comparing genomic and cDNA sequences (open boxes represent exons, circle and arrow represent 5′ and 3′ ends of the isolated cDNAs, respectively). (C) Sequence of mab-21 genomic DNA. The amino acid sequence encoded by the mab-21 open reading frame, defined by cDNA sequences, is shown below in single letter code. The positions of introns are indicated by arrowheads; intronic sequences are not shown. Three mab-21 cDNA species begin with the underlined nucleotides indicated by the letters A, B and C. The putative polyadenylation signal is underlined and the polyadenylation site is indicated by an arrow. The EcoRI site used to generate a frame shift mutation for confirming the identification of mab-21 is labelled.

DNA sequence analysis revealed that the cDNA clones contained an open reading frame encoding a protein of 386 amino acids (Fig. 4C). Search of GenBank and EMBL databases with GCG FASTA and TFASTA identified no previously described proteins with significant similarity to this conceptual protein. Thus the mab-21 locus encodes a novel protein. Furthermore, no amino acid sequence motifs were identified that suggested a possible function or cellular location of the protein.

mab-21 function in ray production is required during late L3 or L4 larval stages

As discussed above, the male tail mutant phenotype of mab-21 suggested that mab-21 was required for a pattern-formation mechanism that directed cell fate choices of seam cells during the late L3 larval stage. To determine whether mab-21 gene function was required at a time consistent with this model, we expressed mab-21 from a heat-shock promoter at various times in a mab-21 mutant background (Fig. 5). We found that the interval during which heat shock resulted in the highest frequency of rescue of the mab-21 mutant phenotype (both the ray 6 defect and the ectopic ray defect) included the time when ray precursor cells were present. Therefore mab-21 function could be required by one or more ray precursor cells. Alternatively, heat shock starting at this time might be necessary in order to accumulate enough mab-21 protein for function at a later time.

Fig. 5.

Rescue of the mab-21 mutant phenotype by expression of a cDNA transgene from a heat-shock promoter. Hours of postembryonic development and larval stages are shown at the left. At the stages of the cell lineage indicated by the dashed lines, animals were placed at 32°C for 3 hours, then returned to 20°C. The number of adult males with fully wild-type tails is shown to the right. R6 is marked by *.

Fig. 5.

Rescue of the mab-21 mutant phenotype by expression of a cDNA transgene from a heat-shock promoter. Hours of postembryonic development and larval stages are shown at the left. At the stages of the cell lineage indicated by the dashed lines, animals were placed at 32°C for 3 hours, then returned to 20°C. The number of adult males with fully wild-type tails is shown to the right. R6 is marked by *.

mab-21 functions cell autonomously for choice of ray identity by R6, but non-autonomously for choice of hypodermal versus neuroblast cell fate by T.apapa

Results of our previous cell ablation experiments with wildtype animals suggested that genes required for specification of ray morphological identities acted autonomously within the terminal branches of the ray lineages (Chow and Emmons, 1994). (Terminal branches are defined as those branches leading to or contributing to single rays.) Therefore we expected that mab-21(+) was required in R6 or a descendant of R6 for correct specification of ray 6 morphology, ultrastructure and position. In order to test this prediction, as well as to determine whether the choice of hypodermal versus neuroblast cell fate by T.apapa was cell autonomous, we carried out a mosaic analysis of mab-21 function, as well as a laser ablation study in a mab-21 mutant background. We found that expression of mab-21(+) within the V6 lineage was both necessary and sufficient for wild-type ray 6 position and morphology, whereas expression of mab-21(+) within T.apapa was not necessary for preventing expression of the neuroblast fate by this cell. Expression within either the T or the V6 lineage was sufficient for preventing expression of the neuroblast cell fate by T.apapa. The laser ablation study confirmed that the presence or absence of R6 can affect the fate of T.apapa.

For mosaic analysis, we utilized a strain carrying the three linked mutations mab-21(bx53)ncl-1(e1865)unc-36(e251) III as well as a free duplication, sDp3(III;f), that carries wild-type alleles of each of these genes (Fig. 4A). sDp3 is spontaneously lost during the cell lineage with a frequency variously estimated to be 1 per 400 cell divisions (Kenyon, 1986) or 1 per 300 cell divisions (Herman, 1989). Mosaic animals were identified as Unc animals with wild-type male tails, or nonUnc animals with mutant male tails, and the probable point of duplication loss was determined for such animals by analysis of the Ncl phenotype of cells representing several lineages (Fig. 6; Materials and Methods). By this means, 36 mosaic animals were identified and could be placed into one of four classes. The cells scored and the points of duplication loss for each mosaic class are shown in Fig. 6; the phenotypes of the mosaics are summarized in Table 2.

Table 2.

V6 and T lineage phenotypes of mosaic animals

V6 and T lineage phenotypes of mosaic animals
V6 and T lineage phenotypes of mosaic animals
Fig. 6.

Mosaic analysis of mab-21. The lineage diagram shows the relationships between V5, V6, T and the cells in which the Ncl phenotype was used to score for the presence of the free duplication. The arrowheads identify the point of loss of the duplication in various classes of mosaics (number in parentheses is the number of animals scored in each class). When the point of duplication loss is ambiguous, the arrowhead points to the last cell in which the loss could have occurred.

Fig. 6.

Mosaic analysis of mab-21. The lineage diagram shows the relationships between V5, V6, T and the cells in which the Ncl phenotype was used to score for the presence of the free duplication. The arrowheads identify the point of loss of the duplication in various classes of mosaics (number in parentheses is the number of animals scored in each class). When the point of duplication loss is ambiguous, the arrowhead points to the last cell in which the loss could have occurred.

Class I represented probable loss of the duplication in AB.p (Ia) or in the AB.pl (Ib) or AB.pr (Ic) branches; these animals most likely lacked a wild-type copy of the gene in T.apapa. Animals of this class had wild-type male tails. This indicated that mab-21(+) gene function was not necessary within descendants of AB.p, which included T and its descendants T.apapa and R7, for either wild-type ray 6 or correct choice of the hypodermal cell fate by T.apapa.

Class II animals had lost the duplication in AB.a (IIa), or in the AB.al (IIb) or AB.ar (IIc) branches; these animals most likely lacked a wild-type gene copy in R6. Such animals were mutant for ray 6 on one (IIb and IIc) or both (IIa) sides, consistent with a requirement for mab-21(+) within R6 or one of its descendants. Expression of mab-21 function within the large hypodermal syncytium, hyp7, appeared unlikely, because Class II mosaics would be expected to have large numbers of syncytial nuclei carrying the mab-21(+) gene from AB.p. Class II animals were wild type for choice of hypodermal cell fate by T.apapa, indicating expression of mab-21(+) in R6 was not necessary for correct specification of T.apapa cell fate.

In class III animals, the duplication was lost in AB and, as expected, such animals were bilaterally mutant for ray 6. Two of the three class III animals had ectopic rays unilaterally. Taken together with the previous results, this indicates that expression of mab-21(+) in either the AB.a or AB.p branches is sufficient for specification of wild-type T.apapa cell fate, but that expression in at least one of these two lineages is necessary. Finally, the single animal of class IV, where the duplication was lost in P1, had a wild-type male tail as expected.

One possible interpretation of the mosaics is that the presence of mab-21(+) activity in the V6 lineage prevents expression of the neuroblast cell fate by a T.apapa cell lacking mab-21 gene function (class I mosaics). This suggested the existence of an interaction between R6 and T.apapa that could affect the fate of T.apapa. We obtained direct evidence for such an interaction by cell ablation experiments carried out on mab-21 mutant animals (Table 3). Consistent with our earlier results in a wild-type background (Chow and Emmons, 1994), most ablations in a mab-21 mutant background had no effect on unablated cells. However, ablation of R6 or its mother reduced the frequency of expression of the neuroblast fate by T.apapa from around 45% to 2.3% (1/43). Therefore, in a mab-21 mutant background, presence of R6 causes T.apapa to express the neuroblast cell fate with increased frequency.

Table 3.

Frequency of T.apapa-derived ray after ablation of seam cells

Frequency of T.apapa-derived ray after ablation of seam cells
Frequency of T.apapa-derived ray after ablation of seam cells

This interaction between R6 and T.apapa in a mab-21 mutant background might be either direct or indirect. R6 itself might send an inductive signal or block an inhibitory signal received by T.apapa (direct interaction), or R6 might simply by its physical presence cause T.apapa to be exposed to an inducing signal or prevent exposure to a blocking signal from another source (indirect interaction). Only the first of these two alternatives is consistent with the results of the mosaic analysis. Expression of mab-21(+) within one or more cells of the AB.a lineage, presumably R6, prevents interaction between R6 and T.apapa, or makes it ineffective (as does expression of mab-21(+) within T.apapa itself). Since gene expression within R6 alters the interaction between R6 and T.apapa, this interaction is likely to be occurring directly between these two cells (Fig. 7).

Fig. 7.

Model for action of mab-21 to block a cell signal. A signal from R6 causes T.apapa to express the ray neuroblast cell fate. Action of mab-21 within either cell blocks the effect of this signal.

Fig. 7.

Model for action of mab-21 to block a cell signal. A signal from R6 causes T.apapa to express the ray neuroblast cell fate. Action of mab-21 within either cell blocks the effect of this signal.

mab-21 mutants affect the differentiation of four cells present bilaterally in the posterior hypodermis of the male tail. Two of these cells, R6st and R6B, are components of ray 6, one of nine sensory rays, while a third (R6.p) is a hypodermal product of the same cell sublineage that generates the cells of ray 6. In mab-21 mutants, each of the three cells R6st, R6B and R6.p adopts a fate or differentiates in a manner similar to that of the corresponding cell of the adjacent more anterior ray sublineage (ray 4). We show that mab-21(+) activity is likely to be required within the ray 6 sublineage. However, we do not know whether its action is required cell autonomously within each ray 6 cell individually. It could be that the effects on some cells are secondary to effects on other cells. One possible explanation for the diverse effects of mab-21 mutations on these three cells is that mab-21(+) activity is required only by the ray precursor cell R6, which in a mab-21 mutant background appears to assume the identity of its anterior neighbor R4. This interpretation is consistent with the earliest possible time of action of mab-21, which was when R6 was present.

mab-21 mutations also affect a fourth cell, a hypodermal seam cell similar to R6. This cell, T.apapa, is born at the same time as R6 and, like R6, moves posteriorly out of the seam after it is born and becomes the immediate posterior neighbor of R6 (Fig. 1C,D). In mab-21 mutants, T.apapa was frequently transformed into a ray precursor cell and expressed the ray sublineage. Thus, in mab-21 mutants, both T.apapa and R6 assume characteristics of more anterior seam cells. It seems most likely that mab-21 acts as part of a pattern-formation mechanism that dictates the fates of seam cells during the late L3 and early L4 larval stages.

We demonstrate that, in a mab-21 background, a signal from R6 causes T.apapa to express the ray neuroblast cell fate. Action of mab-21(+) in either R6 or T.apapa blocks the effect of this signal. This evidence for signalling between seam cells and its modulation by cell-autonomous functions is consistent with previous observations, both in C. elegans and in other organisms. A signal from the posterior seam cell T inhibits the expression of ray lineages by its anterior neighbor V6 (Waring and Kenyon, 1990). In wild type, V6 overcomes this inhibition by the cell-autonomous action of the homeobox transcription factor pal-1 (Waring and Kenyon, 1991). Signalling between seam cells is necessary for expression of the postdeirid neuroblast cell fate by V5.pa (Waring et al., 1992). In this case, effective signalling requires cell contact and this must be maintained on both sides of V5.p in order for this cell to divide asymmetrically (Austin and Kenyon, 1994). Seam cells in C. elegans, like epithelial cells in other organisms, are joined by adherens junctions (Priess and Hirsh, 1986; Baird et al., 1991). Signals passing between epithelial cells that influence their fates are important in the development of many if not all animals (for reviews see Horvitz and Herskowitz, 1992; Greenwald and Rubin, 1992; Peifer et al., 1993).

One possible consequence of signalling between epidermal cells is modulation of the levels of expression of HOM-C/Hox genes, which give cells their regional identities. In C. elegans, the effect of the above-mentioned T inhibition on V6 is proposed to be to prevent expression of the HOM-C/Hox gene mab-5, which is required for the male-specific divisions of the V6 lineage, including the ray sublineage (Waring and Kenyon, 1990, 1991). Likewise, a likely consequence of contact between V5.p and its neighbors is modulation of the expression of mab-5 within the V5 lineage (Austin and Kenyon, 1994).

We have shown previously that the HOM-C/Hox genes mab-5 and egl-5 function in specifying the identities of rays 16 (Chow and Emmons, 1994). mab-5 is most closely related to Drosophila Antennapedia, while egl-5 is most closely related to Drosophila AbdominalB (Wang et al., 1993). The relative levels of expression of these two genes within the terminal cells of the ray lineages help to specify the morphological identity assumed by each ray. In particular, egl-5 is weakly haploinsufficient for expression of the identity of ray 6: in egl-5(0)/+ heterozygotes, ray 6 occasionally (7%) assumes the identity of ray 4 and fuses with ray 4 (Chow and Emmons, 1994).

The function of mab-21 may be related to the role of egl-5 in specification of the development of ray 6. mab-21 mutants are fully penetrant for the same ray 6 to 4 transformation phenotype weakly expressed in egl-5 heterozygotes. Furthermore, mab-21 mutations, themselves recessive, are dominant enhancers of the haploinsufficient phenotype of egl-5. Heterozygosity for a mab-21 mutation increased the frequency of transformation of ray 6 to ray 4 in a heterozygous egl-5 background from 7% to over 30% (Chow and Emmons, 1994). Thus, a decreased level of egl-5(+) gene function makes R6 sensitive to the level of mab-21(+) gene function. Because egl-5 modifies the cellular environment in which mab-21 acts, this argues that the two genes function in the same or related pathways in determination of the identity of ray 6. A third gene also implicated in this pathway is mab-18, the C. elegans homolog of Pax-6 (Zhang and Emmons, 1995). As mab-21 also acts in other body regions in addition to the posterior seam of males, and indeed may have an essential embryonic function, it is possible that mab-21 plays a much wider role in implementing the action of HOM-C/Hox genes or possibly of Pax-6.

In spite of the fact that HOM-C/Hox genes have been recognized for some time as encoding transcription factors that play a key role in regional specialization within the metazoan body, large gaps remain in our understanding of their mode of action. Because the mab-21 gene encodes a putative protein of hitherto unreported amino acid sequence, it is not possible to predict whether it might play a role in regulation of transcription of egl-5 or in target gene selectivity of egl-5, possibly as a cofactor. Another possibility for the function of a HOMC/Hox gene modifier such as mab-21 is that it acts in the upstream pathways that restrict the expression of HOM-C/Hox genes to certain body regions. Thus a further possibility for mab-21 action is as a component of the pattern-formation mechanisms that set the expression levels of HOM-C/Hox genes in the seam cells.

We thank H. Chamberlin and P. Sternberg for the sy155 allele of mab-21, J. Priess for MH27 monoclonal antibody, B. Meyer for TY1353 and J. Miwa for MJ569. We are grateful to A. Leroi for his comments on the manuscript. Most nematode strains used in the mapping studies were received from the Caenorhabditis Genetics Center, which is funded by the NIH National Center for Research Resources (NCRR). K.L.C. was supported by a Martin Foundation Postdoctoral Fellowship. This work was supported by NIH grant R01 GM39353.

Andrew
,
D. J.
and
Scott
,
M. P.
(
1992
).
Downstream of the homeotic genes
.
New Biol
.
4
,
5
15
.
Austin
,
J.
and
Kenyon
,
C.
(
1994
).
Cell contact regulates neuroblast formation in the Caenorhabditis elegans lateral hypodermis
.
Development
120
,
313324
.
Baird
,
S. E.
,
Fitch
,
D. H. A.
,
Kassem
,
I. A. A.
and
Emmons
,
S. W.
(
1991
).
Pattern formation in the nematode epidermis: determination of the arrangement of peripheral sense organs in the C. elegans male tail
.
Development
113
,
515
526
.
Botas
,
J.
(
1993
).
Control of morphogenesis and differentiation by HOM/Hox genes
.
Curr. Opin. Cell Biol
.
5
,
1015
1022
.
Botas
,
J.
,
Moscoso del Prado
,
J.
and
Garcia-Bellido
,
A.
(
1982
)
Gene dose titration analysis in the search of trans-regulatory genes in Drosophila
.
EMBO J
1
,
307
310
.
Brenner
,
S.
(
1974
).
The genetics of Caenorhabditis elegans
.
Genetics
,
77
,
7194
.
Cassada
,
R.
,
Isnenghi
,
E.
,
Culotti
,
M.
and
von Ehrenstein
,
G.
(
1981
).
Genetic analysis of temperature-sensitive embryogenesis mutants in Caenorhabditis elegans
.
Devel. Biol
.
84
,
193
205
.
Chow
,
K. L.
and
Emmons
,
S. W.
(
1994
).
HOM-C genes and four interacting loci determine the morphogenetic properties of single cells in the nematode male tail
.
Development
120
,
2579
2593
.
Devereux
,
J.
,
Haeberli
,
P.
and
Smithies
,
O.
(
1984
).
A comprehensive set of sequence analysis programs for the VAX
.
NAR
12
,
387
395
.
Finney
,
M.
and
Ruvkun
,
G.
(
1990
).
The unc-86 gene product couples cell lineage and cell identity in C. elegans
.
Cell
63
,
895
905
.
Fire
,
A.
,
White-Harrison
,
S.
and
Dixon
,
D.
(
1990
).
A modular set of lacZ fusion vectors for studying gene expression in Caenorhabditis elegans
.
Gene
93
,
189
198
.
Greenwald
,
I.
and
Rubin
,
G. M.
(
1992
).
Making a difference: the role of cellcell interactions in establishing separate identities for equivalent cells
.
Cell
68
,
271
281
.
Hall
,
D. H.
(
1995
).
Electron microscopy and 3D image reconstruction
. In
C. elegans: Modern Biological Analysis of an Organism. Methods in Cell Biology
, Vol.
48
. (
H. F.
Epstein
and
D. C.
Shakes
, eds.)
Academic Press, New York
, pp.
395
436
.
Herman
,
R. K.
(
1984
).
Analysis of genetic mosaics of the nematode Caenorhabditis elegans
.
Genetics
108
,
165
180
.
Herman
,
R. K.
(
1989
).
Mosaic analysis in the nematode Caenorhabditis elegans. J
.
Neurogen
.
5
,
1
24
.
Hodgkin
,
J.
,
Horvitz
,
H. R.
and
Brenner
,
S.
(
1979
).
Nondisjunction mutants of the nematode Caenorhabditis elegans Genetics
91
,
67
94
.
Horvitz
,
H. R.
and
Herskowitz
,
I.
(
1992
).
Mechanisms of asymmetric cell division: two Bs or not two Bs, that is the question
.
Cell
68
,
237
255
.
Kenyon
,
C.
(
1986
).
A gene involved in the development of the posterior body region of C
.
elegans. Cell
46
,
477
487
.
Kramer
,
J. M.
,
French
,
R. P.
,
Park
,
E. C.
and
Johnson
,
J. J.
(
1990
).
The Caenorhabditis elegans rol-6 gene, which interacts with the sqt-1 collagen gene to determine organismal morphology, encodes a collagen
.
Mol. Cell. Biol
.
10
,
2081
2089
.
Kusch
,
M.
and
Edgar
,
R. S.
(
1986
).
Genetic studies of unusual loci that affect body shape of the nematode Caenorhabditis elegans and may code for cuticle structural proteins
.
Genetics
113
,
621
639
.
Levy
,
A. D.
,
Yang
,
Y.
and
Kramer
,
J. M.
(
1993
).
Molecular and genetic analysis of the Caenorhabditis elegans dpy-2 and dpy-10 collagen genes: a variety of molecular alterations affect organismal morphology
.
Mol. Biol. Cell
.
4
,
803
817
.
Liu
,
K. S.
and
Sternberg
,
P. W.
(
1995
).
Sensory regulation of male mating behavior in Caenorhabditis elegans
.
Neuron
14
,
1
20
.
Loer
,
C. M.
and
Kenyon
,
C. J.
(
1993
).
Serotonin-deficient mutants and male mating behavior in the nematode Caenorhabditis elegans
.
J. Neurosci
.
13
,
5407
5417
.
McGinnis
,
W.
and
Krumlauf
,
R.
(
1992
).
Homeobox genes and axial patterning
.
Cell
68
,
283
302
.
Mello
,
C. C.
,
Kramer
,
J. M.
,
Stinchcomb
,
D.
and
Ambros
,
V.
(
1991
).
Efficient gene transfer in C. elegans: Extrachromosomal maintenance and integration of transforming sequences
.
EMBO J
.
10
,
3959
3970
.
Palazzolo
,
M. J.
,
Hamilton
,
B. A.
,
Ding
,
D.
,
Martin
,
C. H.
,
Mead
,
D. A.
,
Mierendorf
,
R. C.
,
Raghavan
,
K. V.
,
Meyerowitz
,
E. M.
and
Lipshitz
,
H.D.
(
1990
).
Phage lambda cDNA cloning vectors for subtractive hybridization, fusion-protein synthesis and Cre-loxP automatic plasmid subcloning
.
Gene
88
,
25
36
.
Peifer
,
M.
,
Orsulic
,
S.
,
Pai
,
L. M.
and
Loureiro
,
J.
(
1993
).
A model system for cell adhesion and signal transduction in Drosophila
.
Development
1993
Suppl.
,
163
-
176
.
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
.
Rosenbluth
,
R. E.
,
Cuddeford
,
C.
and
Baillie
,
D. L.
(
1985
).
Mutagenesis in Caenorhabditis elegans. II. A spectrum of mutational events induced with 1500 R of gamma-radiation
.
Genetics
109
,
493
511
.
Schierenberg
,
E.
,
Miwa
,
J.
and
von Ehrenstein
,
G.
(
1980
).
Cell lineages and developmental defects of temperature-sensitive embryonic arrest mutants in Caenorhabditis elegans
.
Dev. Biol
.
76
,
141
159
.
Simon
,
M. A.
,
Bowtell
,
D. D.
,
Dodson
,
G. S.
,
Laverty
,
L T. R.
and
Rubin
,
J. M.
(
1991
).
Ras1 and a putative guanine nucleotide exchange factor permorm crucial steps in signaling by the sevenless protein tyrosine kinase
.
Cell
67
,
701
716
.
Sulston
,
J. E.
and
Horvitz
,
H. R.
(
1977
).
Post-embryonic cell lineages of the nematode Caenorhabditis elegans
.
Dev. Biol
.
56
,
111
156
.
Sulston
,
J. E.
,
Albertson
,
D. G.
and
Thompson
,
J. N.
(
1980
).
The Caenorhabditis elegans male: postembryonic development of nongonadal structures
.
Dev. Biol
.
78
,
542
576
.
Sulston
,
J.
and
Hodgkin
,
J.
(
1988
).
Methods
. In
Wood
,
W. B.
(ed.)
The Nematode Caenorhabditis elegans
.
Cold Spring Harbor Laboratory Press
. pp
587
606
.
Wang
,
B. B.
,
Muller-Immergluck
,
M. M.
,
Austin
,
J.
,
Robinson
,
N. T.
,
Chisholm
,
A.
and
Kenyon
,
C.
(
1993
).
A homeotic gene cluster patterns the anteroposterior body axis of C
.
elegans. Cell
74
,
29
42
.
Waring
,
D.
and
Kenyon
,
C.
(
1990
).
Selective silencing of cell communication influences anteroposterior pattern formation in C. elegans
.
Cell
60
,
123
131
.
Waring
,
D. A.
and
Kenyon
,
C.
(
1991
).
Regulation of cellular responsiveness to inductive signals in the developing C. elegans nervous system
.
Nature
350
,
712
715
.
Waring
,
D. A.
,
Wrischnik
,
L.
and
Kenyon
,
C.
(
1992
).
Cell signals allow the expression of a pre-existent neural pattern in C. elegans
.
Development
116
,
457
466
.
Zhang
,
Y.
and
Emmons
,
S. W.
(
1995
).
Specification of sense organ identity by a C. elegans Pax-6 homolog
.
Nature, in press
.