[
Worm Breeder's Gazette,
1986]
During a sabbatical year at the MRC Laboratory in Cambridge, England, I isolated mutants that affect formation of the copulatory structures of the adult male tail (mab), and studied the process of tail morphogenesis. The most obvious feature of the adult male tail is the cuticular fan with its nine rays. This structure, consisting of a large fold in the outer layer of the cuticle, is formed when the cells that inhabit the posterior region of the L4 larva retract just before the L4/adult molt (Sulston et al., Dev. Biol. 78, 542-576 (1980)). As the cells withdraw anteriorly, the outer layer of the newly-formed adult cuticle is left behind and eventually folds in a precise way to form the fan. Two questions arise concerning this process: what is the mechanism of cell withdrawal, and how is the cuticular folding accomplished in such a precise way, with reproducible and sharp boundaries? It can be seen, by comparison of the volume of the animal before ar_ after withdrawal of the posterior cells, that retraction must be accompanied by extrusion of a large amount of fluid from the body. This fluid passes into the growing space between the retreating cells and the L4 cuticle, and is under pressure. In order to see whether the pressure affected the morphogenetic movements, a hole was made in the L4 cuticle with the laser microbeam during the morphogenetic process. When the cuticle was punctured, the fluid escaped in a rush, and morphogenesis stopped. The partially retracted cells bulged back against the L4 cuticle. This result could mean that retraction of the posterior cells is driven by pressure between the L4 and adult cuticles. Alternatively, it might simply mean that the newly-forming adult cuticle is not yet strong enough to resist internal body pressure. To test this idea worms were observed in hypertonic medium ( e.g., 6% glucose), where pressure is reduced in both the space in the tail between the L4 and adult cuticles, and in the body generally, because of loss of water through the cuticle. In hypertonic medium, morphogenesis of the tail is also arrested. This result makes it appear that liquid pressure in the space between the adult and L4 cuticles is necessary per se for the morphogenetic movement. This pressure would act along with retraction caused by muscles in the tail, and possibly other cellular mechanisms, to mold the final shape of the adult body. As retraction proceeds, folding of the fan occurs against the taut L4 cuticle, which may act as a sort of scaffold or mold for the process. This may help to explain the preciseness of the shape of the fan, but it cannot account for its boundaries. Retraction of the body occurs well beyond the boundaries of the fan. Beyond the fan the inner and outer layers of the adult cuticle no longer separate. Instead, the outer cuticular layer adheres to the inner layer and the surface of the body. The outer cuticular layer and its behavior can be observed under Nomarski optics during the morphogenetic process. A critical aspect of fan formation must involve synthesis of a special type of cuticle, in which inner and outer layers are not joined, over the surface where the fan will eventually form during retraction. Electron micrographs of serial sections of males in the late L4 stage were examined to determine what cells might be responsible for making this special cuticle. The electron micrographs were made by Sulston and coworkers in their earlier studies of the male. Much of the tale hypodermis in the late L4 larva is made up from the posterior daughters of the ray precursor cells, Rn (n=1-9), which come to lie roughly underneath the region where the outer cuticular layer will separate to form the fan. Further analysis of electron micrographs, along with laser ablation experiments, are necessary to determine whether these and possibly other hypodermal cells are indeed responsible for making the fan. In genetic studies, nine new mab mutations were isolated by screening males in clones of a mutagenized him strain for abnormalities. One mutation is a new allele of the previously- identified gene
mab-11, and the others are probably in new mab genes. Abnormalities affect various structures, from the spicules to the fan and rays, and ranged from subtle to severe. One severe mutant was examined in more detail, and surprisingly was found to undergo a nearly normal morphogenetic process, up until the time that the L4 cuticle was breached in molting. At this point, the normally-shaped tail bulged back out and spoiled the nearly-normal fan, in somewhat the same manner as occurred when the L4 cuticle was prematurely punctured with the laser. This gene therefore appears to encode a product required to hold the definitive fan in place. The same mutant is ts for alae formation in the adults of both sexes, and has severe cuticle defects in the L1 larva, which is osmotically sensitive. Therefore the protein may also be used to maintain alae ( present exclusively in L1's, dauers, and adults - dauers were not examined in these experiments). Alae, like the fan, consist of a cuticular fold, and are formed by seam cells. (The ray precursor cells, Rn, are descended from seam cells.) We propose that this mutation affects a cuticle component required to maintain the structure of both alae and fan, and may identify a component of a morphogenetic 'subprogram' used in making these two structures.
[
Worm Breeder's Gazette,
1982]
We have isolated and studied the properties of the DNA sequence responsible for the Bristol-Bergerac polymorphisms discovered in our earlier work (Emmons et al., PNAS 76, 1333, 1979). We have found that these polymorphisms arise because the Bergerac genome contains about 200 copies of a 1.7kb transposon inserted at dispersed sites in the DNA, whereas the Bristol genome contains only about 20 copies of the same sequence. We conclude that this sequence is a transposon from the fact that at each site of insertion, in both Bristol and Bergerac, the 1.7kb segments have precisely the same organization. The insertion events have therefore been site-specific with respect to the inserted DNA, but not (highly) site-specific with respect to the target, and it is this property which is characteristic of the transposons that have been studied in other organisms such as bacteria, Drosophila, and yeast. By analogy with the nomenclature adopted in yeast (Ty1), we have named the 1.7kb DNA element Tc1. The arrangement of Tc1 is (probably) different in various wild C. elegans strains we have looked at (see following report). We have further found that Tc1 is not stably inserted, but appears to excise rapidly from three sites of insertion in Bergerac we have studied, lending further support to the conclusion that this is a motile DNA sequence. We isolated Tc1 by screening a Bergerac clone bank in lambda 1059 with a Bristol BamHI fragment that hybridized to polymorphic fragments on a genomic Southern. Similar experiments were carried out in David Hirsh's laboratory using a Dictyosteleum discoideum actin probe ( Newsletter Vol. 7, p.51). So far by several criteria the two inserts isolated by Hirsh's group and by us appear to be identical. The conclusions that Tc1 has a conserved organization and is found dispersed among diverse genomic sequences come from an analysis of Southern hybridizations of genomic DNA using the cloned Tc1 as a probe. When enzymes that do not cut within Tc1 are used a smear of hybridization is seen in Bergerac and a series of bands in Bristol. When enzymes that cut within Tc1 are used, bands due to internal fragments and smears due to the variable external sites are seen, and all the patterns are consistent with the conclusion that every genomic element is 1.7kb in length and has the same internal restriction sites as our cloned copy. By measuring the intensity of hybridization to the internal fragments we concluded that Bergerac has around 200 copies of Tc1 and Bristol around 20. We observe excision of Tc1 when non-Tc1 sequences lying on either side of a site of Tc1 insertion are used as a probe in a genomic Southern. In this case the polymorphic Bristol and Bergerac restriction fragments at the insertion site are seen, the Bergerac fragment being 1.7 kb larger than the Bristol one. In all Bergerac DNA preparations we have examined, we see in addition a small amount of the smaller fragment lacking Tc1. This is true for probes specific for three different Tc1 insertion sites, indicating the excision is likely to be a property of the element and not of the site. To find out how rapidly the excision occurs, we examined separately 8 DNA preparations from grandchildren and great grandchildren of 8 single Bergerac worms. All 8 show the same low level of excision, which we estimate to be about 1%. Excision is therefore very rapid, probably occurring in every worm. To find out whether excision occurs in the germ line, we have been cultivating Bergerac continuously for the past 6 months. DNA from these worms will be examined for complete loss of Tc1 at individual sites. Tc1 is the most highly repeated non-ribosomal sequence in the Bergerac genome and is the only highly repeated sequence present in different amounts in Bristol and Bergerac DNA. We have shown this by hybridizing nick-translated, whole genomic DNA to Bristol and Bergerac clone banks in lambda1059., and comparing the results with those obtained using cloned Tc1 or rDNA as a probe. The only plaques that show a difference with Bristol and Bergerac genomic probes are those that hybridize as well to Tc1. The number of such plaques in the two clone banks is consistent with there being around 200 copies of Tc1 in Bergerac and 20 in Bristol. Finding this transposon has stimulated us to spend a considerable amount of time attempting to demonstrate hybrid dysgenesis in crosses between Bristol and Bergerac. The large number of copies of Tc1 in Bergerac made it seem likely that this strain should be used as the male parent in such crosses. Since Bergerac males were said to be infertile, we first mated Bristol males to Bergerac hermaphrodites. The hybrid sons of this cross were then mated to marked Bristol hermaphrodites (Dpy, E61). Most F1 progeny of this cross are normal. Hermaphrodites give normal broods with the correct proportion of Dpy worms and males are fertile. However, 3 hermaphrodite worms (out of over 100 examined) gave very abnormal broods. They segregated dead embryos in large numbers. Progeny that hatched often died before maturity, or were sterile; 20% were males, and 13% were intersexes of unstable genotype. Hermaphroditic progeny that did mature gave abnormal broods with a similar spectrum of phenotypes. We have not obtained such results with reciprocal crosses, but the numbers of worms involved is too low to draw any conclusions from this. Attempts to increase the frequency of abnormal broods by various means have not been successful. No worms with stable mutant phenotypes were obtained.
[
Worm Breeder's Gazette,
1983]
The excision events that Tc1 elements in Bergerac undergo at high frequency occur in somatic cells and not in germ cells. We drew this conclusion earlier from experiments showing that Tc1 elements were stably inherited during long-term propagation. We have now been able to demonstrate directly that excision is confined to somatic cells by studying the amount of excision that has occurred as a function of age in staged populations of worms. We synchronized Bergerac cultures by isolating embryos from young, gravid hermaphrodites with hypochlorite. These embryos were allowed to develop and at various stages samples were taken for DNA isolation. In each DNA preparation the amount of excision that had taken place at three Tc1 sites was assessed using flanking-sequence probes. (We used pCe1001, isolated in our laboratory, pCe(Br)T1, isolated in Hirsh's laboratory, and pCe1003, a subclone from phage lambda 3-8 isolated by Bill Sharrock.) For all three probes the fraction of empty sites was found to steadily increase during larval development from as little as 0.6% in embryos to as much as 10% in late larvae. The fraction in adults was lower, 2-3%, as expected from growth of germ cells and internal embryos. Most particularly, however, the level in embryos of the next generation fell once again to the lowest level, indicating that empty Tc1 sites were not inherited. Preliminary results with worms arrested by starvation suggest that excision occurs at a steady rate independent of developmental stage, since the amount of excision was found to correlate with time since fertilization rather than with stage. We have been able to detect extrachromosomal copies of Tc1 that may be the excised elements. When Bergerac DNA is fractionated on an agarose gel without treatment with a restriction endonuclease, and hybridized on a Southern to a Tc1-specific probe, three low molecular weight species can be seen with mobilities that are consistent with a 1.6 kb linear, relaxed circle, and supercoil. The level of hybridization to these species in DNA from L4 larvae corresponds to about one copy per cell, and appears to increase during development, suggesting a relationship to the excision process. We have isolated these molecules on sucrose gradients and begun to analyze their structure with restriction endonucleases. Results so far confirm that the relaxed circle and linear are free copies of Tc1. The linear species has unique (not permuted) ends that correspond to the ends of an inserted Tc1 element. Although failing to identify a Tc1 at the
unc-22 site, our analysis of the recombinants has revealed two putative
unc-22 linked Tc1 polymorphisms, one between
daf-14 and
unc-22 (a 2 m.u. interval), and one between
unc-22 and
dpy-4 (a 4.5 m.u. interval). The putative polymorphism to the left of
unc-22 is a 2.1Kb HindIII fragment containing Tc1 which we have cloned into lambda 590. At present we are trying to isolate a unique sequence segment from this HindIII fragment so that we can use it to position the polymorphism with greater certainty. If the fragment is suitably close to
unc-22 we may be able to use it to walk to the
unc-22 gene. In conclusion, we feel that Tc1 is unlikely to be responsible for generating the
unc-22 mutations, but linked Tc1 polymorphisms may be useful in providing us with a molecular probe for the region.
[
Worm Breeder's Gazette,
1978]
During the past year we have initiated a study of the DNA of C. elegans with a number of long range objectives in mind. We would like to isolate the DNA from genetically defined regions of the genome in order to construct physical maps to go along with genetic maps. We would like to use isolated fragments of DNA as hybridization probes for studies of transcription. If the size of an isolated restriction fragment differs in two strains of the worm, e.g., Bristol and Bergerac, this size can be used as a phenotype to map the genetic location of the restriction fragment. In this way we hope to locate on the genetic map genes, such as the ribosomal genes, which can be physically isolated but in which mutations have not yet been identified. The size of restriction fragments can also be used as a sensitive method to search for changes in the primary structure of the DNA during development. Here we report our initial progress in these experiments. Isolation of nucleic acids We have found that worms can be completely dissolved and digested by proteinase K in 1% SDS at 65 C, allowing isolation of DNA of high molecular weight and poly-A-containing RNA which is active in an in vitro translation system. Worms, usually frozen in liquid N2 and ground with a mortar before melting, are taken up in .1M tris pH 8.5, . 05M EDTA, .2M NaCl, 1% SDS. (Freezing and grinding is probably not necessary.) Proteinase K (EM Laboratories, Inc.; available from Scientific Products) is added to 200 lambda/ml and the mixture is heated to 65 C for 15' with occasional gentle rocking to mix. During this time the mixture clears almost completely and all worm carcasses disappear. The highly viscous solution is then extracted three times with phenol and once with chloroform-isoamyl alcohol (24:1). DNA may be separated from RNA at this point by precipitating nucleic acids with ethanol and winding out the DNA. We further purify DNA from RNA by digestion with RNase followed by phenol extraction and ethanol precipitation. Unfortunately, worms, particularly adults, contain a particulate material (a polysaccharide?) which copurifies with DNA through organic extractions and ethanol precipitations. This material results in blue DNA solutions, and may be responsible for the indigestibility of some DNA preparations with restriction endonucleases. Much of this material can be removed by spinning the DNA solution at 20,000 rpm for half an hour, and we do this routinely. To obtain C. elegans DNA rigorously free of E. coli DNA, we allow hypochlorited eggs to hatch into M-9 buffer and then purify DNA from the hatched L1's. C. elegans ribosomal DNA DNA coding for 18s and 28s ribosomal RNA (rDNA) can be purified from the bulk of worm DNA as a high density (50% G+C) satellite on cesium chloride gradients (Sulston and Brenner, Genetics 77, 95-104, 1974). The ribosomal genes are tandemly repetitious, containing about 50 copies of each gene. Digestion of this rDNA-containing satellite with restriction endonucleases Bam HI or Sal I gives a single band, the ribosomal unit repeat of 6800 base pairs. The appearance of only one band indicates that the rDNA contains a rather homogeneous repeat, and is the only repetitive DNA in the satellite. This band hybridizes labelled ribosomal sequences at a level 50-100 fold greater than expected for a unique sequence. We have cloned the 6800 base pair Bam fragment thereby providing a probe for hybridization experiments. We have mapped a number of restriction sites within the ribosomal repeat unit. Total worm DNA is digested with the appropriate restriction enzyme(s), run on a 0.7% agarose gel and blotted onto nitrocellulose filter by the technique of Southern. The filter is hybridized to either iodinated 125I-rRNA or to nick-translated 32P- cloned rDNA to identify the restriction fragments containing rDNA. The restriction map is consistent with a homogeneous 6800 base pair unit repeat. Heterogeneous repeat units present in only one copy would not be seen in this analysis, but are currently being looked for. The approximate location of 18s and 28s genes and of spacer regions has been located on the map. Hybridization to Southern blots from heavily loaded gels of digested DNA show a few minor bands, which are presumably fragments from the ends of the tandem repeat, containing some overlap into non-ribosomal DNA. It is also possible that some minor bands are heterogeneous unit repeats. Cloning and characterization of these fragments is in progress. It is striking that the length of the unit repeat is smaller than that found in almost all other eukaryotes. This could be the result of small genes or of very short spacer regions. We have sized the C. elegans large and small rRNAs by electrophoresis of glyoxal-denatured RNA on agarose gels (McMaster and Carmichael, PNAS 74, 4835-4838, 1977) obtaining values of 1700 and 3350 nucleotides, smaller than other eukaryotic rRNAs. In an attempt to locate the ribosomal genes on the genetic map, and to study the inheritance of repetitive DNA, we have compared the restriction cutting patterns of N2 rDNA with those of other strains of C. elegans. Any difference in cutting pattern (most likely due to spacer differences) would be a phenotype, easily mappable. In a comparison of N2 with C. elegans var. Bergerac (J. Brun), rDNA cutting patterns were identical with each of 12 restriction enzymes used. Using several restriction enzymes, the rDNA cutting pattern was also the same with a strain of C. elegans isolated from the wild (D. Russel). rDNA from C. briggsae (B. Zuckerman) did give differences in restriction cutting patterns. The restriction map is similar to that of N2, although the unit repeat is 400 base pairs longer and a few cutting sites are added or deleted. Some fragments appear to be the same in both species. Unfortunately, attempts by us (as well as by Nigon and Dougherty, J. Exp. Zool. 112, 485-503, 1949) to cross C. elegans and C. briggsae have not succeeded. Work at establishing a genetic system for rDNA is continuing. Repeated sequences in C. elegans DNA Sulston and Brenner (Genetics 77, 95-104, 1974) have shown that the DNA of C. elegans contains repetitive components similar to those found in other eukaryotic organisms: namely, inverted repeats, highly repetitive sequences, moderately repetitive sequences, and uniclue sequences. We have undertaken the further characterization of these sequences. So far we have completed initial experiments on the inverted repeat sequences and the moderately repetitive sequences. Inverted repeats. We have studied inverted repeat sequences by electron microscopy and find them to be similar in every way to those found in other eukaryotic organisms. Inverted repeats are visualized by simply melting high molecular weight DNA and spreading it for electron microscopy using the formamide technique of Davis, Simon and Davidson (Methods in Embryology, Vol. XXI, p. 413, 1971). The inverted repeat sequences are then seen as double-stranded stems, or stems with loops at their end, sticking out from the largely single- stranded DNA. Of 37 inverted repeats visualized, those without terminal loops (78%) had stems with a number average length of 250 bp, and those with loops (22%) had stems with a number average length of 340 bp. The number average length of the loops was 800 bp. The inverted repeats appear to be located in clusters in the DNA. Clusters contain a few (3-6) inverted repeats separated by about a thousand base pairs, and clusters are separated from each other by 10 to more than 70 thousand base pairs of DNA containing no inverted repeats. Moderately repetitive sequences. Moderately repetitive sequences are sequences present in the range of 10 to 100 times in the genome. In most eukaryotic organisms about half of such sequences consist of short (300 bp) stretches of repetitive DNA surrounded by unique sequences, and a large fraction of the unique DNA is interspersed in this way, at about one thousand base pair intervals, with moderately repetitive sequences. A few organisms (Drosophila, Chironomous, honey bee, and Achyla--a water mold) lack these short, interspersed repetitive sequences. We have studied the interspersion pattern of repetitive DNA in C. elegans by reassociation kinetics and find that, like Drosophila and the others of the minority group, it appears to lack the highly interspersed component of repetitive DNA. We have compared the rate of reassociation of fragments averaging 300 (120-650) and 2000 (1000- 4000) base pairs in length, using hydroxyapatite binding to assay formation of double strands. DNA of L1's were used after labeling by nick-translation. For shearing, reassociation, and hydroxyapatite binding, the methods of Britten, Graham, and Neufeld (Methods in Enzymology, 29, 363, 1974) have been followed. Seventy-six percent of the 2000 base pair fragments reannealed at a rate expected for unique fragments of that length. This represents only a slight increase over the fraction of the 300 base pair fragments which carry repetitive DNA, an increase from 20% to 24%. This result is consistent with a lack of highly interspersed repetitive DNA. We are presently analyzing the length of moderately repetitive sequences by electron microscopy to determine whether any short repetitive sequences are present at all. Studies on cloned fragments of C. elegans DNA We have constructed a small clone bank of C. elegans restriction fragments. We have used the Bam restriction endonuclease and have inserted the fragments into the pBR313 driver plasmid. Recombinant DNA work with C. elegans DNA is at the P2-EK1 level. We will be happy to share recombinant plasmids. We are using the cloned fragments as hybridization probes to study restriction fragments in worm DNA. A restriction digest of whole- genome DNA is fractionated on an agarose gel and transferred to a millipore filter (a 'Southern transfer'). A plasmid containing a particular cloned fragment is then labeled by nick-translation and hybridized to the filter to reveal the fragments in the whole digest which carry sequences homologous to those of the cloned fragment. We have been analyzing the patterns produced in this way to answer a number of questions: 1. Are the patterns consistent with the arrangement of repetitive DNA determined by COT analysis; that is, do most fragments consist solely of unique sequences? 2. Are there any differences in the patterns given by germ-line and somatic-line DNAs? 3. Are there any differences in the patterns given by Bristol and Bergerac DNAs? These could be used for mapping. Are there any differences between C. elegans and C. briggsae patterns? 4. Can differences in these patterns be used to find cloned fragments that come from genetically defined regions, for example, from regions covered by deletions? By hybridizing 0.1 g of a plasmid nick-translated to more than 10+E7cpm/ g to a filter carrying a digest of a few micrograms of worm DNA we can detect unique fragments after an overnight exposure. We use flashed film and intensifying screens and expose the film at -70 C. We have found that hybridizations at low temperature (e.g., 32 C) in 50% formamide and without Denhardt's solution are convenient and work very well. Thirteen recombinant plasmids (with inserts ranging in size from 1, 000 to 18,000 base pairs) have been hybridized to filters carrying digests of DNA from N2 L4 hermaphrodites, N2 L1 hermaphrodites, Bergerac L1 hermaphrodites, and C. briggsae (mixed population). All hybridize to a fragment in N2 DNA equal in size to the cloned insert they carry, indicating that no rearrangements have taken place during cloning. Nine plasmids hybridize to several (up to 10) additional bands. Even most inserts of less than 2000 base pairs (5 out of 8) hybridize to more than one band. From the COT analysis described earlier we would expect 76% of such fragments to consist entirely of unique DNA. Whether these figures are inconsistent, and if so, why, remains to be seen. We have used L4 hermaphrodites as a source of 'germ line' DNA in these experiments. By comparing DNA from them to DNA from newly hatched L1's, which lack a gonad, we can search for restriction fragments present in the germ line but absent from the somatic line. No such fragments have been found; so far the L1 and L4 patterns are identical. We have also started to use DNA isolated from sperm nuclei (a gift from Michael Klass), which will allow a much more rigorous comparison of germ and somatic line sequences. (In addition we are hoping that a comparison of sperm and hermaphrodite DNAs will allow, by examining the relative intensities of bands, identification of fragments from the X-chromosome.) Comparison of the bands in Bergerac and Bristol DNA's shows that these DNAs are not identical. Five Bristol bands (including two of the cloned inserts) appear to have a different size in Bergerac; that is, they are missing in the Bergerac pattern and one new band is present. We would like to find out whether these differences are due to single base changes or to rearrangements. This degree of difference between these strains suggests that genetic mapping by restriction fragments is feasible. Comparison of the C. elegans patterns with those of C. briggsae shows (to our surprise) that these DNAs are highly diverged. None of the 13 cloned fragments is present unaltered in the briggsae genome, and in fact 9 hybridize to no fragment whatsoever in briggsae DNA. Since we expect (but have not checked) that the proteins of these ( almost indistinguishable) worms would be very similar, this raises the possibility that DNA sequences present in both species are coding sequences.
[
Worm Breeder's Gazette,
1990]
The dramatic morphogenesis of the fan and rays at the end of the L4 larval stage of the C. elegans male raises several questions. How is the reproducible shape of the fan determined? Why does the outer layer of the adult cuticle that makes up the fan fold at such precise boundaries? What causes the tips of the rays to attach at specific sites either in the dorsal or ventral surfaces, or at the margin of the fan? Mutations that cause fusions of rays and alteration of ray attachment sites have identified genes required for ray formation (see Baird & Emmons, Newsletter Vol 11 #2,
p116, 1990). What is the function of such genes? Studies of the topography of hypodermal cells in the tail during the L3 and L4 have provided some answers to these questions. Hypodermal cell boundaries in males of various ages have been visualized by means of monoclonal antibody MH27, kindly provided by R. Waterston and coworkers. This antibody is presumed to react with belt desmosomes that ring each hypodermal cell and join it to its neighbors. A central role in fan formation appears to be played by the tail seam (SET), a syncytium formed by fusion of the posterior daughters of the five ray precursor cells R1-R5 (Figure panels a and b). In the adult male the SET is a narrow extension of the body seam, to which it is attached by a desmosome (Sulston et al., Dev. Biol 78, 542-576, 1980). We find that earlier in the L4 the SET is much enlarged and covers the surface that will become the dorsal surface of the fan ( Figure panels c and d). The boundary between the SET and the ventral hypodermis (
hyp7, or possibly R6.p or a cell descended from T) runs where the margin of the fan will form, and could define this fold- point. The future position of the margin can be determined from the arrangement of the ray tips, as discussed below. The positions of the ray tips in the adult fan appear to be defined by distinct affinities of (probably) the ray structural cells for the two hypodermal domains defined by the SET and the ventral hypodermis. The cells of the rays are born in the early L4 in the hypodermis ( Figure panel b). By mid L4 a single cell remains on the surface for each ray, presumably the structural cell, the two neurons having sunk into the body. The structural cells arrange themselves in a specific way with respect to the boundary between the SET and the ventral hypodermis (Figure panel d). Structural cells of rays that attach to the dorsal surface of the fan (rays 1, 5, and 7) move up into the SET. Those of rays that attach to the ventral surface of the fan (rays 2, and 4; rays 8 and 9 have not been visualized) move downwards into the ventral hypodermis. The structural cell of ray 3, which extends right to the margin of the fan in the adult, remains at the boundary between SET and ventral hypodermis. The arrangement seen with the antibody stain is congruent with the pattern of ray tips visible with Nomarski optics late in the L4 at the beginning of morphogenesis (Figure panel e). The arrangement shown in Figure panel d is in fact still uncertain for rays 5 through 9. It is presented as best we can make it out from our data, and by analogy with rays 1-4. The above observations make it possible to interpret mutations that cause fused rays as resulting in adjacent ray structural cells that lie in the same hypodermal domain. This might be because structural cells have lost or changed identities, or because they cannot express specific affinities for the SET or ventral hypodermis. If structural cells are not segregated into separate hypodermal domains, they may come together and become surrounded by a single sheet of hypodermis. Our evidence for this interpretation is preliminary. In
mab-20, which causes fusions of ray 1 to ray 2, and ray 3 to ray 4, compound rays run right to the margin of the fan in the adult, and ray tips have altered positions consistent with the structural cells lying between the SET and the ventral hypodermis (Figure panel f). In
mab-18, which causes transformation of ray 6 to a ray 4-like ray, usually fused to ray 4, the tip of ray 6 is found in the position of and adjacent to the tip of ray 4. We have not yet studied mutants with the MH27 antibody. [See Figure 1]
[
Worm Breeder's Gazette,
1997]
The
mab-18 gene product is a putative transcription factor consisting of the homeodomain plus the carboxy-terminal, serine-threonine-rich transcriptional activation domain of the C. elegans Pax-6 homolog (1).
mab-18 is one part of a complex locus that also includes the gene
vab-3.
vab-3 mutations primarily affect morphogenesis of the head (2).
mab-18 is necessary for expression of the unique identity of ray 6 in the male tail. In
mab-18 mutants, ray lineages are unaffected, but cells of ray 6 form a ray at the positon where ray 4 normally forms, and usually co-assemble with ray 4 cells to form a large, compound ray. This phenotype is interpreted as a ray 6 to ray 4 identity transformation. We have studied the expression pattern of MAB-18 by indirect immunofluorescence staining with affinity purified anti-MAB-18 antibodies (from a serum raised against a GST-MAB-18 fusion protein). We find that MAB-18 first appears in the ray 6 precursor cell R6. Staining is localized to the cytoplasm, and is relatively excluded from the nucleus. Cytoplasmic staining, and apparent nuclear exclusion, persist through the two rounds of cell division of the ray sublineage. After completion of the sublineage, when the ray cells begin to differentiate into ray neurons and support cell, staining disappears in the cytoplasm and appears in the nucleus, where it remains into adulthood. In our previous studies with a
mab-18 reporter gene, we unexpectedly found the reporter was expressed during the ray sublineage of ray 8, a ray not known to be affected by
mab-18 mutations. Expression of the reporter in ray 8 cells did not persist into the adult, but ended after the sublineages were completed. With the anti-MAB-18 antibodies, we find cytoplasmic staining in the ray 8 precursor cell, R8, and also in the ray 7 precursor cell, R7. As for ray 6, cytoplasmic staining persists through these sublineages. However, when the sublineages are completed, staining disappears altogether, and does not appear in the nucleus. In several other non-ray lineages where the antibodies detect MAB-18 or VAB-3, we observed only nuclear staining. We speculate that prior cytoplasmic accumulation of MAB-18 and sudden release into the nucleus might be important for the rapid differentiation of ray 6 cells. Alternatively, or in addition, regulation of nuclear entry might help direct action of MAB-18 to a single ray. One regulatory system might confine transcription of the
mab-18 gene to the three rays 6, 7, and 8, while a second regulatory system triggers nuclear entry of the MAB-18 protein only in ray 6. Recently, Mann and Abu-Shaar (3) have found that nuclear entry of the homeodomain protein extradenticle in Drosophila embryonic midgut requires both dpp and wingless signals from visceral mesoderm. Thus regulated nuclear entry of homeodomain transcription factors may be a widespread phenomenon. 1: Zhang, Y., and Emmons, S.W. (1995) Nature 377, 55-59. 2: Chisholm, A.D., and Horvitz, H.R. (1995) Nature 377, 52-55. 3: Mann, R.S., and Abu-Shaar, M. (1996) Nature 383, 630-633.