The Phenotype of mes-2, mes-3, mes-4 and mes-6, Maternal-Effect Genes Required for Survival of the Germline in Caenorhabditis elegans, Is Sensitive to Chromosome Dosage

Corresponding author: Susan Strome, Department of Biology, Indiana University, Bloomington, IN 47405, sstrome{at}bio.indiana.edu (E-mail).

Communicating editor: R. K. HERMAN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mutations in mes-2, mes-3, mes-4, and mes-6 result in maternal-effect sterility: hermaphrodite offspring of mes/mes mothers are sterile because of underproliferation and death of the germ cells, as well as an absence of gametes. Mutant germ cells do not undergo programmed cell death, but instead undergo a necrotic-type death, and their general poor health apparently prevents surviving germ cells from forming gametes. Male offspring of mes mothers display a significantly less severe germline phenotype than their hermaphrodite siblings, and males are often fertile. This differential response of hermaphrodite and male offspring to the absence of mes⁺ product is a result of their different X chromosome compositions; regardless of their sexual phenotype, XX worms display a more severe germline phenotype than XO worms, and XXX worms display the most severe phenotype. The sensitivity of the mutant phenotype to chromosome dosage, along with the similarity of two MES proteins to chromatin-associated regulators of gene expression in Drosophila, suggest that the essential role of the mes genes is in control of gene expression in the germline. An additional, nonessential role of the mes genes in the soma is suggested by the surprising finding that mutations in the mes genes, like mutations in dosage compensation genes, feminize animals whose male sexual identity is somewhat ambiguous. We hypothesize that the mes genes encode maternally supplied regulators of chromatin structure and gene expression in the germline and perhaps in somatic cells of the early embryo, and that at least some of their targets are on the X chromosomes.

THE germline is crucial for the perpetuation of species. In metazoans, it is set apart from the soma relatively early in development, and must be specified to follow its unique developmental program of extensive proliferation, meiosis, and the production of gametes (for review see MARSH and GOODE 1994 ). The task of identifying factors that participate in specifying the germline and initiating its developmental program can be approached genetically by screening for mutants defective in early aspects of germline development. In organisms that rely heavily on a maternal contribution of RNA and proteins, the crucial factors are likely to be encoded by maternal-effect genes. This logic prompted us to carry out extensive screens for maternal-effect sterile mutants in the nematode, Caenorhabditis elegans (CAPOWSKI et al. 1991 ). This paper describes analysis of the major class of mutants that came from those screens.

In C. elegans, the germline (P4 cell) is set apart from the soma by the 16–24-cell stage of embryogenesis. The primordial germ cell divides only once during embryogenesis. Extensive proliferation of the germline occurs during larval development, resulting in expansion of the germline from two cells in first stage larvae (L1s) to ~1500 cells in adult hermaphrodites and ~500 cells in adult males. Meiosis commences in the L3 stage, sperm are formed in the L4 stage, and in hermaphrodites oocyte production begins in the adult stage; males continue to make sperm throughout adulthood (see KEMPHUES and STROME 1997 and SCHEDL 1997 for reviews of germline establishment and development).

To identify genes encoding early-acting factors required uniquely for germline development, we screened for maternal-effect sterile (mes) mutants (CAPOWSKI et al. 1991 ). The phenotype of homozygous mes worms is dependent on the genotype of their mother: mes hermaphrodites produced by mes/+ mothers are fertile, whereas mes hermaphrodites produced by mes mothers are sterile. Even mes/+ hermaphrodites produced by mating mes mothers to wild-type males are sterile, revealing that maternal mes⁺ product is both necessary and sufficient for normal germline development. Mutations in four of the five genes identified in this screen (mes-2, mes-3, mes-4, and mes-6) result in a common mutant phenotype: the hermaphrodite offspring of mes mothers undergo apparently normal germline development during embryogenesis but develop into sterile adults with a 10–100-fold reduction in the number of undifferentiated germ cells and no gametes (CAPOWSKI et al. 1991 ; PAULSEN et al. 1995 ).

As described in this paper, we see a dramatic difference in the severity of the Mes phenotype when comparing male and hermaphrodite siblings produced by mes mothers. In contrast to the severely affected germline seen in mutant hermaphrodites, mutant male siblings generally possess a well-proliferated and healthy-appearing germline, produce hundreds of sperm, and are fertile. We present evidence that the difference in expressivity of the Mes phenotype in hermaphrodites and males depends on X chromosome dosage and not on the sexual phenotype of the worms. Molecular analysis has revealed that two of the mes genes encode homologs of Polycomb group (Pc-G) proteins in Drosophila (R. HOLDEMAN, S. NEHRT and S. STROME, unpublished results; I. KORF, R. HOLDEMAN, Y. FAN and S. STROME, unpublished results). Pc-G proteins repress gene expression and are thought to act by controlling chromatin structure. A role for MES proteins in modulation of chromatin and repression of gene expression in the germline is consistent with the sensitivity of the mutant phenotypes to chromosome dosage. In addition to their essential role in the germline, the mes genes may serve a nonessential role in somatic cells of the early embryo. This is revealed by our unexpected finding that mes mutations can affect the sexual fate of worms, probably through an effect on dosage compensation in the early embryo.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

General techniques and strains:
Maintenance and genetic manipulation of C. elegans were carried out as described by BRENNER 1974 . Unless stated otherwise, experiments were conducted at room temperature (20–22°). The permissive and restrictive temperatures for all temperature-sensitive mutations used in our experiments were 16° and 25°, respectively. Most strains were provided by the Caenorhabditis Genetics Center, including the tetraploid (4n) stock (strain SP346). fog-1(q253ts) was from JUDITH KIMBLE. tra-1(e1099) was from BARBARA ROBERTSON and WILLIAM WOOD. tra-1(e1575gf) was from TIM SCHEDL. sdc-3(y52) was from BARBARA MEYER. C. elegans strain N2 variety Bristol was the wild-type parent of all nematode strains used. Mutations, deficiencies, and balancers used are listed by linkage group (LG):

LG I: mes-3(bn35), fog-1(q253ts), unc-11(e47), dpy-5(e61), dpy-14(e188), glp-4(bn2ts), hDp20 (I;V,f)

LG II: mes-2(bn11, bn27, bn76), tra-2(e1095), dpy-10(e128), unc-4(e120), mnC1, rol-1(e91)

LG III: tra-1(e1099, e1575gf ), plg-1(e2001), unc-32(e189), unc-79(e1068), ced-4(n1162), glp-1(q175), eT1 (III,V)

LG IV: mes-6(bn64, bn66), him-8(e1489), fem-3(q20ts,gf ), unc-26(e205), DnT1[unc(n754)let] (IV;V)

LG V: mes-4(bn58, bn67), dpy-11(e224), sdc-3(y52, y128), unc-76(e911), DnT1[unc(n754)let] (IV;V)

mes mutations were maintained as balanced stocks or in trans to appropriately marked chromosomes. mes-2 mutations were balanced with mnC1 (HERMAN 1978 ). mes-3 mutations were maintained as homozygotes in combination with hDp20 to provide mes-3⁺ function, or in heterozygotes in trans to one of the following recessively marked chromosomes: unc-11 dpy-5 or dpy-5 dpy-14. mes-4 and mes-6 mutations were balanced with the reciprocal translocation DnT1, which has a dominant Unc and a recessive lethal phenotype (E. FERGUSON, personal communication).

Nomenclature:
Standardized C. elegans genetic nomenclature (HORVITZ et al. 1979 ) has been used. Genes and transcripts are given three-letter, lowercase, italicized designations, e.g., mes-2. Protein products are given nonitalicized, all uppercase designations e.g., MES-2. The mutant phenotypes caused by mutating a particular gene are given the three-letter designation with only the first letter being uppercase, e.g., Mes. In this paper, the term bal will be used to designate any of the balancer chromosomes or marked chromosomes described in the previous section, and Bal will be used to describe the phenotype of each.

Staining nuclei:
To visualize nuclei in intact animals, single worms were placed in 1 µl of water on a polylysine-coated slide (MILLER and SHAKES 1995 ), and the slide was briefly waved over the flame of an alcohol burner until all the water had evaporated and the worms were dry. Eight µl of an equal volume mixture of self-hardening mounting medium (Gelutol; Monsanto, St. Louis, MO; MILLER and SHAKES 1995 ) and Hoechst Stain (40 mM NaC1, 10 mM Tris-Cl, pH 7.4, 1 mM EDTA, pH 8.0, 20% glycerin, 10 µg/ml Hoechst 33342) were put on an 18 x 18-mm coverslip and placed on the specimen.

To visualize nuclei in extruded gonads, worms were placed in a 10-µl drop of Solution A [100 µg/ml Hoechst 33342, 0.5x egg salts (EDGAR 1995 ), and 0.25 µM levamisole to promote hypercontraction] on a polylysine-coated slide. After allowing the levamisole to take effect, worms were cut open with a scalpel just posterior to the pharynx or anterior to the anus. The released gonads (and other tissues) were incubated in Solution A for 2–3 min to allow the Hoechst dye to penetrate. Then 10 µl of Solution B (5% glutaraldehyde, 0.5x egg salts, 2 µg/ml Hoechst 33342) were added. Samples were covered with a coverslip and viewed immediately.

All samples were viewed on an Axioskop (Carl Zeiss, Thornwood, NY) equipped with Nomarski differential interference contrast and epifluorescence optics. Photographs were taken with Tri-X film and developed with D-76 developer (Eastman Kodak, Rochester, NY).

Synchronizing worms:
Gravid hermaphrodites were cleaned of Escherichia coli by allowing them to crawl on agar plates lacking bacteria. Worms were incubated in 20 µl of distilled water at room temperature for 6–10 hr, and then cut in half with a scalpel to release their embryos. The embryos were put in a humid chamber overnight. By the next day, all the embryos had hatched and arrested as L1s. L1s were transferred to plates spread with bacteria and allowed to grow to the desired age.

Laser ablation of distal tip cells:
A general description of laser microsurgery in C. elegans is described elsewhere (BARGMANN and AVERY 1995 ). Worms were mounted on 0.4% agarose pads containing 10 mM sodium azide and overlaid with coverslips. Ablations were performed in hermaphrodites at the late L2/early L3 stage of development using a VSL-337 laser (Laser Science Inc., Newton, MA) and coumarin dye (AVERY and HORVITZ 1987 ). Only one distal tip cell (DTC) was ablated per worm, leaving the other gonad arm to serve as a control for the behavior of the germline in the presence of the DTC. We monitored the effectiveness of our technique by ablating the DTC of wild-type hermaphrodites and demonstrating that all germ cells became sperm. After operating, worms were placed on individual plates, allowed to develop for 48 hr, Hoechst stained as described above, and assayed for the number of sperm, meiotic germ nuclei, and/or undifferentiated germ nuclei present. Results were scored blindly, without knowledge of the genotype of the specimen or which DTC had been destroyed.

Construction and analysis of mes; glp-4 mutants:
mes/bal; glp-4 strains were created with the following mes alleles: mes-2(bn11), mes-4(bn58), mes-4(bn67), and mes-6(bn66). Crosses were done at 16° unless otherwise noted. (P0) glp-4 males (raised at permissive temperature) were mated to mes/bal hermaphrodites. (F1) Progeny were allowed to mate with their siblings. The desired matings were between mes/+; glp-4/+, and bal/+; glp-4/+ F1s to produce mes/bal; glp-4 F2s. (F2) L4 hermaphrodites from these matings were picked to individual plates and allowed to produce self progeny for 24 hr. For the mes-4 and mes-6 constructions, Unc L4s (containing DnT1) were picked. After 30–50 F3s had been laid, the F2s were transferred to new plates. The first set of plates was moved to 25°, and the second set was kept at 16°. (F3) Worms raised at 25° were scored for the presence of 100% Glp sterile worms. For the mes-2 construction, the plates were also scored for the presence of mnC1 worms. (F4) 16° plates corresponding to broods with Glp steriles at 25° were scored for the presence of ~33% Mes sterile worms. This identified new strains that were mes/bal; glp-4.

To perform experiments with the mes-2/mnC1; glp-4 strain, individual non-Dpy non-Unc hermaphrodites were allowed to lay eggs for 24 hr at 16° and were then transferred to 25°. mes-2/mnC1; glp-4 hermaphrodites produced fertile progeny at 16°, including Dpy Uncs. Their germlines were non-Mes non-Glp-4. The siblings raised at 25° were non-Mes Glp-4. mes-2; glp-4 hermaphrodites produced all sterile progeny at 16°, with no Dpy Uncs. Their germlines were Mes non-Glp-4. The siblings raised at 25° were Mes Glp-4.

To perform experiments with the mes-4/DnT1; glp-4 and mes-6/DnT1; glp-4 strains, individual Unc (containing DnT1) and non-Unc (lacking DnT1) hermaphrodites were allowed to lay eggs for 24 hr at 16° and were then transferred to 25°. The mes/DnT1; glp-4 hermaphrodites produced fertile progeny at 16°. Their germlines were non-Mes non-Glp-4. The siblings raised at 25° were non-Mes Glp-4. mes; glp-4 hermaphrodites produced all sterile progeny at 16°. Their germlines were Mes non-Glp-4. The siblings raised at 25° were Mes Glp-4.

Worms were harvested as L3s or young adults. The age of the worms was confirmed by examination of ventral hypodermal and vulval cells under Nomarski optics (SULSTON and HORVITZ 1977 ). Worms were stained with Hoechst to assess the appearance of mutant germ nuclei.

Construction and analysis of mes; glp-1 mutants:
mes/bal; unc-32 glp-1/++ stocks were created using mes-2(bn11), mes-2(bn27), mes-4(bn58), mes-4(bn67), and mes-6(bn69). (P0) unc-32 glp-1/eT1 males were crossed to mes/bal hermaphrodites. (F1) Outcross male progeny (unc-32 glp-1/++ or eT1/+, in combination with mes/+ or bal/+) were mated en masse to new mes/bal hermaphrodites. (F2) L4 hermaphrodites were picked to individual plates, allowed to produce self progeny, and scored for the presence of the desired phenotypic classes: F3 Unc Glp sterile worms, F3 Bal worms, and F4 Mes sterile worms. (If DnT1 was the balancer chromosome used, then the dominant phenotype allowed its presence to be identified and balanced worms to be selected from among the F2). The resulting mes/bal; unc-32 glp-1/++ stocks were maintained by cloning fertile mes/bal (or mes/mes) worms from broods that also contained Unc Glp worms. Experiments were performed by comparing the germlines of non-Mes Unc Glp worms from mes/bal; unc-32 glp-1/++ mothers to the germlines of Mes Unc Glp worms from mes; unc-32 glp-1/++ mothers. Germlines were analyzed in intact young adults that were stained with Hoechst.

Construction and analysis of mes; ced-4 mutants:
We constructed mes/+; unc-79 ced-4/++ strains using mes-2(bn11), mes-4(bn67), and mes-6(bn66). (P0) N2 males were mated to unc-79 ced-4 hermaphrodites. (F1) Heterozygous male cross progeny were mated to mes/bal hermaphrodites. (F2) Non-Bal L4 hermaphrodite offspring were placed on individual plates and allowed to produce self progeny. (F3) Broods were scored for the presence of Unc worms and the absence of Bal worms, indicating that the F2 parent was the desired heterozygote: mes/+; unc-79 ced-4/++. F3 Unc Ced and non-Unc non-Ced hermaphrodites were placed on individual plates. Approximately one fourth of each class was expected to be mes/mes. (F4) Broods were scored for fertility or sterility. We compared the extent of germline proliferation between Mes Unc Ced F4s (produced by mes; unc-79 ced-4 F3s) and Mes non-Unc non-Ced F4s (produced by mes; unc-79 ced-4/++ or mes; ++ F3s). Germlines were analyzed in intact young adults stained with Hoechst.

Construction and analysis of mes; him-8 mutants:
mes-2(bn11) and mes-3(bn35) strains, in the configuration mes/bal; him-8, were made and analyzed as follows: (P0) him-8 males were crossed to mes/bal hermaphrodites. (F1) Male progeny (mes/+; him-8/+ or bal/+; him-8/+) were mated to outcross hermaphrodite siblings (mes/+; him-8/+, genotype was confirmed by allowing the worms to lay eggs before mating them and then analyzing the self progeny). (F2) L4 hermaphrodites were placed on individual plates and allowed to produce two generations of progeny. (F3) Plates containing Bal worms and males were retained. (F4) The desired strain was identified by the presence of Mes sterile hermaphrodites. To compare the germlines of mes; him-8 males and hermaphrodites, non-Bal L4 hermaphrodites were isolated and allowed to produce self progeny. Approximately one third of the non-Bal hermaphrodites were mes; him-8, and they produced all sterile Mes hermaphrodite progeny and semifertile Mes male progeny.

mes-4(bn67)/DnT1; him-8/DnT1 was made by crossing him-8 males to mes/DnT1 hermaphrodites (P0). (F1) mes/+;him-8/+ male progeny were crossed to mes/DnT1 hermaphrodites. (F2) Individual L4 Unc (containing DnT1) hermaphrodites were allowed to produce two generations of progeny. (F4) The desired strain contained Mes sterile hermaphrodites and males. We also constructed a mes-6 him-8 recombinant chromosome and balanced it over DnT1 to create mes-6(bn66) him-8/DnT1. For analysis, Mes males and hermaphrodites were obtained from unmated, fertile, non-Unc (mes; him-8) hermaphrodites.

Mes siblings of each sex were raised at 20°, harvested between 68 and 74 hr after hatching, stained with Hoechst to visualize DNA, and then examined.

Construction and analysis of mes-4; fog-1 mutants:
We constructed mes-4/DnT1; fog-1 strains using mes-4(bn58) and mes-4(bn67), and a fog-1 male stock that was established from a rare spontaneous male. (P0) fog-1 males (raised at permissive temperature) were crossed to mes-4/DnT1 hermaphrodites. (F1) Outcross males were mated to their hermaphrodite siblings, also at 16°. (F2) Unc (containing DnT1) L4 hermaphrodite offspring were placed on individual plates at 16° and allowed to lay ~30 embryos. Then the F2s were transferred to new plates at 25°. (F3) The 25° F3 progeny were scored. Broods consisting of 100% Fog steriles were noted, and the corresponding 16° plates were retained at 16° for one more generation. (F4) The 16° plates were examined for the presence of Mes sterile worms. Strains were maintained at 16° by cloning fertile Unc hermaphrodites.

Fertile non-Unc (mes-4; fog-1) hermaphrodites were transferred from 16° to 25° to examine the feminized germline of their offspring. Fertile non-Unc hermaphrodites were also maintained at 16° to generate Mes offspring without a feminized germline. Fertile Unc (mes-4/DnT1; fog-1) hermaphrodites were transferred from 16° to 25° to verify that non-Mes germlines were feminized by fog-1. Worms were harvested as young adults and stained with Hoechst.

To examine Mes males with a feminized germline, fertile non-Unc (mes-4; fog-1) hermaphrodites were transferred individually from 16° to 25° and mated to fog-1 males (raised at permissive temperature). Outcross mes-4/+; fog-1 males were recovered for analysis. These were compared to mes-4/+; fog-1/+ males, which were obtained by mating mes-4; fog-1 hermaphrodites and N2 males. Worms were harvested as young adults and stained with Hoechst.

Construction and analysis of mes-4; fem-3(g f ) mutants:
We created mes-4/DnT1; fem-3(gf)/DnT1 strains using mes-4(bn58), mes-4(bn67), and a plg-1; fem-3(gf); him-5 strain. The plg-1 mutation causes males to deposit a gelatinous plug at the vulva of their mating partners, and the him-5 mutation causes hermaphrodites to produce ~33% XO male offspring. The plg-1 and him-5 mutations were not retained in the mes-4/DnT1; fem-3(gf)/DnT1 strains. (P0) plg-1; fem-3(gf); him-5 males (raised at permissive temperature) were crossed to mes-4/DnT1 hermaphrodites at 16°. (F1) Non-Unc (lacking DnT1) F1 males were mated to new mes-4/DnT1 hermaphrodites at 16°. (F2) L4 Unc (containing DnT1) hermaphrodites were placed on individual plates at 16° and left there for two generations. (F4) From plates containing Mes non-Unc sterile worms, fertile Unc hermaphrodites were transferred individually to 25°. (F5) Those that produced Fem-3(gf) non-Unc sterile and non-Fem-3(gf) Unc fertile worms identified the desired mes-4/DnT1; fem-3(gf)/DnT1 construct.

Fertile non-Unc [mes-4; fem-3(gf)] hermaphrodites were transferred from 16° to 25° to examine the masculinized germline of their offspring. Fertile non-Unc hermaphrodites were also maintained at 16° to generate Mes offspring without a masculinized germline. Fertile Unc [mes-4/DnT1; fem-3(gf)] hermaphrodites were transferred from 16° to 25° to verify that non-Mes germlines were masculinized by fem-3(gf). Worms were harvested as young adults and stained with Hoechst.

Construction and analysis of mes-4; tra-1 mutants:
We created a mes-4(bn67)/DnT1; tra-1(e1099)/+ strain using the following procedure: (P0) On one plate, tra-1 XX males were crossed to mes-4/DnT1 hermaphrodites. On a second plate, wild-type XO males were crossed to mes-4/DnT1 hermaphrodites. (F1) From the first mating, Unc hermaphrodite progeny were placed individually onto new plates and allowed to lay ~30 F2 embryos. [The purpose of this step was to identify F1 outcross hermaphrodites (DnT1/+; tra-1/+) that would produce one fourth tra-1 male F2s.] Each F1 was then transferred to a new plate and mated to multiple non-Unc (mes/+) F1 males from the second P0 mating plate. (F2) Self progeny of F1 hermaphrodites were scored for the presence of Tra males. From corresponding F1 mating plates, L4 F2 Unc hermaphrodites were picked onto new plates and allowed to produce two generations of progeny. (F3) Broods that segregated Tra males were retained. (F4) The presence of Mes sterile hermaphrodites in this generation identified the desired strain: mes-4/DnT1; tra-1/+. To maintain the stock, Unc hermaphrodites were picked individually from plates that segregated males.

Experiments were performed as follows. Non-Unc L4 hermaphrodites (either mes-4; tra-1/+ or mes-4; +) were separated from males to prevent mating. Hermaphrodites that were fertile at maturity (i.e., offspring of mes-4/DnT1 mothers) were placed onto individual plates. mes-4; tra-1/+ hermaphrodites produced all sterile progeny, one quarter of which were Tra males. We compared the germlines of these Tra Mes males to the germlines of their non-Tra Mes hermaphrodite siblings. We also examined Tra non-Mes control males produced by mes-4/DnT1; tra/+ hermaphrodites. All worms were harvested as young adults and stained with Hoechst.

Construction and analysis of mes-4; tra-2 mutants:
We made a mes-4(bn67)/DnT1; tra-2/+ strain using the following procedure. (P0) Fertile mes males (from a mes-4/DnT1 male stock) were crossed to tra-2/dpy-10 unc-4 hermaphrodites. (F1) Outcross males (mes-4/+; tra-2/+ or mes-4/+; dpy-10 unc-4/++) were mated en masse to mes-4/DnT1 hermaphrodites. (F2) From successful F1 matings, L4 Unc F2s were placed onto individual plates and allowed to produce two generations of progeny. (F3) Plates with Tra males were retained. (F4) The presence of Mes sterile hermaphrodites identified the desired strain: mes-4/DnT1; tra-2/+. Experiments were performed by placing fertile non-Unc hermaphrodites (either mes-4; tra-2/+ or mes-4; +) onto individual plates. The former produced ~25% Mes Tra males. We harvested the Mes Tra males and their Mes non-Tra hermaphrodite siblings as young adults, stained them with Hoechst, and compared their germlines. We also examined Tra non-Mes males produced by mes-4/DnT1; tra-2/+ mothers.

Construction and analysis of mes-4; tra-1(g f) mutants:
A strain containing the dominant tra-1(e1575gf) allele (HODGKIN 1987 ) was maintained by mating tra-1(gf)/+ XX females to wild-type XO males. Offspring included tra-1(gf)/+ XX females (self-sterile, but fertile when mated), tra-1(gf)/+ XO females (self-sterile, but slightly fertile when mated), +/+ XX hermaphrodites (self-fertile), and +/+ XO males (fertile). Potential tra-1(gf)/+ XX and XO females, along with +/+ XX hermaphrodites, were separated from male siblings as L4s. This step was performed to prevent worms from mating with males. In particular, it was necessary to prevent tra-1(gf)/+ XX females from mating because the production of embryos would make the females indistinguishable from +/+ XX hermaphrodites. Upon reaching maturity, tra-1(gf)/+ sterile females were classified as being XX or XO based on size and tail morphology: tra-1(gf)/+ XX females are the size of wild-type hermaphrodites and have a normal spike tail; tra-1(gf)/+ XO females are smaller and have a truncated, stumpy tail.

We generated mes-4(bn67); tra-1(gf)/+ XO hermaphrodites using the following procedure: (P0) tra-1(gf)/+ XX females were mated to mes-4 XO males (derived from a mes-4/DnT1 male stock). (F1) Outcross mes-4/+; tra-1(gf)/+ XX females (identified by examining the tail morphology of sterile worms that had been separated from males at the L4 stage) were mated to new mes-4 XO males. (F2) Outcross XX females [mes-4; tra-1(gf)/+ or mes-4/+; tra-1(gf)/+] were identified and then mated individually to new mes-4 XO males. (F3) Broods that were all sterile identified mothers that were mes-4; tra-1(gf)/+. Stumpy-tailed individuals were presumed to be mes-4; tra-1(gf)/+ XO females, and spike-tailed worms were presumed to be either mes-4; tra-1(gf)/+ XX females or mes-4; + XX hermaphrodites. Young adults were harvested and stained with Hoechst. We compared the germlines of mes-4; tra-1(gf)/+ XO females, mes-4; tra-1(gf)/+ XX females, mes-4; + XX hermaphrodites, and mes-4; + XO males.

Construction and analysis of 2X:3A mes mutants:
Tetraploid stocks fall into two classes (MADL and HERMAN 1979 ): (1) hermaphrodites having four sets of autosomes and four X chromosomes (4X:4A) produce ~99% hermaphrodite self progeny, and (2) hermaphrodites having four sets of autosomes but only 3X chromosomes (3X:4A) produce 25-40% male progeny. Male offspring of 3X:4A hermaphrodites are usually 2X:4A animals (X:A = 0.5). Triploid mes mutants were generated by mating 2X:4A tetraploid males to 2X:2A diploid mes mothers. The resulting outcross progeny were 2X:3A mes/+/+ mutants. All mes alleles used in this analysis were marked with a visible mutation. As a control, we also mated tetraploid males to diploid hermaphrodites that carried the marker mutation but were mes⁺. The following mes mutations and markers were used: mes-2(bn27) marked with unc-4, mes-3(bn35) marked with dpy-5, and mes-4(bn67) and mes-6(bn64), both marked with dpy-11. 2X:3A outcross progeny were scored for sexual phenotype by light microscopy and by staining with Hoechst. Some of the 2X:3A outcross progeny were able to produce gametes, perhaps as a result of the extra copies of the mes⁺ alleles from the tetraploid parent. The presence of oocytes allowed us to confirm that the worms were indeed triploids: oocytes contained six bivalents and six monovalents instead of the wild-type number of six bivalents (MADL and HERMAN 1979 ).

Construction and analysis of mes; sdc-3(y52) mutants:
sdc-3(y52) was derived from the strain sdc-3(y128)/sdc-3(y52) unc-76. sdc-3(y128) disrupts dosage compensation, causing homozygous XX mutants to be Dpy. sdc-3(y52) disrupts sex determination, resulting in the transformation of homozygous XX mutants into males. Because these two sdc-3 mutations affect different processes, they can be maintained as a trans-heterozygous stock (KLEIN and MEYER 1993 ). The mes mutations combined with sdc-3(y52) were mes-2(bn76) marked with rol-1, mes-3(bn35) marked with dpy-5, and mes-4(bn67) and mes-6(bn66) marked with dpy-11.

(P0) Wild-type males were crossed to sdc-3(y128)/sdc-3(y52) unc-76 hermaphrodites. (F1) Individual male outcross were mated to single mes marker/bal hermaphrodites. (F2) L4 hermaphrodite progeny from successful mating plates were cloned and allowed to produce self progeny. (F3) Broods were scored for the presence of homozygous mes marker worms (Rol or Dpy) and homozygous sdc-3(y52) unc-76 worms (Unc). mes marker; sdc-3(y52) unc-76/++ hermaphrodites were placed on separate plates and allowed to produce progeny. (F4) All of the F4s were Mes steriles, and ~25% were also Unc males [sdc-3(y52)]. (The unc marker was scorable in a rol or dpy background.) We counted the number of Mes Sdc-3(Tra) worms present and determined their sexual phenotype using both light microscopy and Hoechst staining.

mes-6 and dpy-11, which are unlinked, were separated after the F1 cross. At subsequent steps, homozygous mes-6 mutants were identified by the presence of all sterile progeny. Since sdc-3 and mes-4 are ~1 map unit apart on LGV, a recombination event was required to create the double mutant. Of 60 F2 hermaphrodites cloned, three recombinants were identified.

To control for the potential effects of the marker-bearing chromosomes, we also analyzed marker; sdc-3(y52) unc-76 animals produced by marker; sdc-3(y52) unc-76/++ hermaphrodites. The parent hermaphrodites were obtained from the mes marker; sdc-3(y52) unc-76/++ strains described above, resulting either from recombination between the mes and marker genes (mes-2, mes-3, and mes-4) or from independent assortment of the mes gene from the marker gene (mes-6). Homozygous marker mutants (mes⁺) were identified by the production of all fertile progeny.

Statistical analysis:
Descriptions of all statistical analyses can be found in AMBROSE and AMBROSE 1987 . The rank sum test, a nonparametric test for significant differences between two samples of equal or unequal size, was used to compare the average number of germ nuclei in different populations of worms. To select between two methods of calculating t, the F test was applied to establish whether or not the variances of the samples were significantly different. These tests were performed on the data presented in Figure 2, Figure 6, Figure 7, and Figure 8. Results are significant at the 0.05 level.

View larger version (73K): [in this window] [in a new window]   Figure 1. —Germlines from mes-4+ and mes-4 mutant hermaphrodites and males. The panels show fluorescence micrographs of him-8 and mes-4(bn67); him-8 mutant gonads that were dissected from adult worms and stained with Hoechst 33342 to visualize DNA (see MATERIALS AND METHODS). The mutant gonads shown are representative of the more proliferated gonads observed among mes-4 hermaphrodites and males. (A) Hermaphrodite gonad arms. The him-8 gonad contains many evenly spaced germ nuclei that are progressing through various stages of mitosis (near the distal tip, indicated by an arrow), meiosis, and oogenesis (indicated by an o); the region of the gonad containing sperm is not shown. The mes-4; him-8 gonad contains many fewer germ nuclei that are enlarged and unevenly spaced. (B) Male gonads. Both gonads contain many evenly spaced nuclei at various stages of mitosis (near the distal tip, indicated by an arrow), meiosis, and spermatogenesis (sperm are indicated by an s). Bar, 10 µm.

View larger version (26K): [in this window] [in a new window]   Figure 2. —Mutations in ced-4 do not suppress the germline proliferation defects associated with mutations in mes-2, mes-4, and mes-6. Adult worms were stained with Hoechst to visualize DNA, and their germ nuclei were counted (see MATERIALS AND METHODS). The histograms compare the average number of germ nuclei in Unc (black bars) and non-Unc (gray bars) offspring of mes; unc-79 ced-4/++ mothers; standard deviations are shown. Between 13 and 29 worms were scored for each genotype. For each mes allele, the average number of germ nuclei in mes hermaphrodites does not differ significantly from the average number in mes; unc-79 ced-4 hermaphrodites (see MATERIALS AND METHODS for statistical analysis).

View larger version (75K): [in this window] [in a new window]   Figure 3. —Germ nuclei in young adult mes-2; glp-4 hermaphrodites have a Mes-2, not a Glp-4, morphology. Sterile worms were stained with Hoechst to visualize DNA. (A) glp-4 worms raised at a restrictive temperature (25°). (B) Offspring of mes-2(bn11); glp-4 mothers raised at a permissive temperature (16°). (C) Offspring of mes-2(bn11); glp-4 mothers raised at a restrictive temperature. Arrows point to undifferentiated germ nuclei. The germ nuclei in glp-4 single mutants raised at a restrictive temperature appear to be arrested in prophase I of mitosis, and they contain distinct condensed chromosomes. Germ nuclei in mes-2; glp-4 mutants are large and diffuse, resembling those in mes-2 single mutants. Bars, 5 µm.

View larger version (23K): [in this window] [in a new window]   Figure 4. —Tests of mes-2, mes-4, and mes-6 germ cell behavior in the absence of DTC signaling. (A) In the absence of glp-1+ function, mes mutant germ cells fail to proliferate but do not differentiate. Bars show the average number of undifferentiated germ cells in adult hermaphrodites; standard deviations are shown. Numbers above each bar indicate the percentage of worms with gametes or gamete-like structures. (Gray bars) unc-32 glp-1 offspring of mes/+; unc-32 glp-1/++ mothers. All germ cells differentiated into sperm. (White bars) mes; ++ or mes; unc-32 glp-1/++ offspring of mes; unc-32 glp-1/++ mothers. Most germ cells remained undifferentiated or died. A few oocyte-like and/or sperm-like structures were observed in some worms bearing weak mes alleles. (Black bars) mes; unc-32 glp-1 offspring of mes; unc-32 glp-1/++ mothers. Most germ cells remained undifferentiated or died, but there were a few sperm-like structures in 5% of mes-4(bn58) worms and 7% of mes-6(bn66) worms. Worms represented by the open and black bars were siblings. Between 13 and 25 worms were scored for each genotype. Standard deviations for mes-4(bn58) and mes-6(bn66) germlines were ±231 and ±248, respectively. (B) Ablation of the DTC in mes-4 and mes-6 mutant worms reduces germ cell proliferation and does not induce gamete formation. The DTC of one gonad arm was ablated at the L3 stage. The unablated gonad arm served as a control for the level of germline proliferation and gamete production in mutant worms. Each horizontal pair of bars shows the two gonad arms of an individual worm scored 48 hr after microsurgery: white bars show the unoperated gonad arm, and black bars show the arm in which the DTC was ablated. Asterisks indicate that a few sperm were present.

View larger version (26K): [in this window] [in a new window]   Figure 5. —Effects of the him-8 genetic background on germline proliferation in mes mutant hermaphrodites. Histogram bars show the average number of germ nuclei in synchronized populations of hermaphrodite offspring produced by homozygous mes (gray bars) and mes; him-8 (black bars) mothers; standard deviations are shown. Between 13 and 30 worms were scored for each genotype.

View larger version (27K): [in this window] [in a new window]   Figure 6. —Mutations in all the mes genes cause more severe germline defects in XX hermaphrodites than in XO males. Sibling hermaphrodites and males produced by homozygous mes; him-8 mothers were stained with Hoechst to visualize and count germ nuclei. Between 12 and 30 worms were scored for each genotype. Each bar represents the percentage of worms containing a similar number of undifferentiated germ nuclei, within a range of 150 germ nuclei for hermaphrodites and 50 germ nuclei for males. Thus, each panel shows the frequency distribution of worms containing different numbers of germ nuclei. The percentage of worms that contained gametes or gamete-like structures is indicated at the top of each panel. For each mes mutant, the average number of germ nuclei in mes; him-8 XX hermaphrodites is significantly different from the average number in mes; him-8 XO males.

View larger version (17K): [in this window] [in a new window]   Figure 7. —The phenotype of mes-4(bn67) is more severe in XX than XO worms, regardless of the sex of the worms. Adult worms were stained with Hoechst to visualize and count germ nuclei. Between 28 and 48 worms were scored for each genotype. Each panel shows a frequency distribution of worms containing different numbers of germ nuclei (see legend to Figure 6). (A) Comparison of sibling mes-4; + or mes-4; tra-2/+ XX hermaphrodites and mes-4; tra-2 XX males produced by mes-4; tra-2/+ XX mothers. Both XX hermaphrodites and XX males had severely underproliferated germlines. (B) Comparison of sibling mes-4; + XO males and mes-4; tra-1(gf)/+ XO stumpy-tailed females produced by a mating between mes-4; tra-1(gf)/+ XX tapered-tail females and mes-4; + XO males. Both XO males and XO females had relatively well-proliferated germlines. The average number of germ nuclei does not differ significantly between different populations of XX worms (Mes hermaphrodites and Mes Tra-2 males) or between different populations of XO worms [Mes males and Mes Tra-1(gf) females]. However, the difference is significant between different populations of hermaphrodites/females [Mes XX hermaphrodites and Mes Tra-1(gf) XO females] and between different populations of males (Mes XO males and Mes Tra-2 XX males).

View larger version (35K): [in this window] [in a new window]   Figure 8. —Increasing the dose of X chromosomes enhances the germline defects caused by mes mutations. Comparison of the germlines in adult XO males (white bars), XX hermaphrodites (gray bars), and XXX Dpy hermaphrodites (black bars) produced by mes; him-8 XX mothers. Worms were stained with Hoechst to visualize and count germ nuclei. Between 11 and 44 worms were scored for each genotype. Each bar represents the average number of germ nuclei per worm. For each mes allele, the average number of germ nuclei is significantly different between mes; him-8 XO, XX, and XXX worms. Numbers above the bars indicate the percent of worms that contained gametes or gamete-like structures. Gametes in mes-2; him-8, mes-3; him-8, and mes-6; him-8 males consisted of sperm only. Although sperm were not counted, there were hundreds of sperm in most males. Gametes in mes-4; him-8 males included what appeared to be endomitotic oocytes in addition to, or instead of, sperm. This putative feminization of some mes-4 XO germlines has not been analyzed further. Gametes in XX hermaphrodites were few in number (usually less than 5) and included endomitotic oocytes, normal-appearing oocytes (with six condensed bivalents), and/or sperm (which sometimes looked normal and sometimes looked abnormal).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Germ cells in mes mutant hermaphrodites are unheal-thy and prone to degeneration:
In wild-type hermaphrodites, germ nuclei in the mitotic region are similar in size and uniformly arranged around the periphery of each syncytial gonad arm. In mes mutant hermaphrodites, germ nuclei are reduced in number compared to wild type, appear swollen, and show an abnormal arrangement in the gonad; the cytoplasm also appears "coagulated" (CAPOWSKI et al. 1991 ; PAULSEN et al. 1995 ). Figure 1A shows wild-type and mes-4 gonad arms that were released from adult hermaphrodites and stained with the DNA dye Hoechst 33342. The reduction in number and the altered appearance and arrangement of germ nuclei in the mes-4 gonad arms are apparent.

For most mes alleles, late larval and adult hermaphrodites show a decrease in the number of germ cells relative to earlier larval stages (CAPOWSKI et al. 1991 ; PAULSEN et al. 1995 ), indicating that at least some germ cells die. Programmed cell death is a normal fate choice for many somatic cells and some germ cells in wild-type hermaphrodites (HENGARTNER 1997 ). PAULSEN et al. 1995 previously tested whether programmed cell death is responsible for the decline in the number of germ cells in mes-3 mutant hermaphrodites. Mutations in mes-3 were combined with a mutation in ced-3 or ced-4. Mutations in either of these two ced genes prevent the programmed cell deaths that occur during normal development (ELLIS and HORVITZ 1986 ). Mutations in the ced genes did not suppress the germline death observed in mes-3 mutants, indicating that germline death in mes-3 worms is not caused by inappropriate activation of the programmed cell death pathway. To extend this analysis to the other mes genes, we constructed mes; unc-79 ced-4/++ hermaphrodites and examined the germlines of their homozygous mes; unc-79 ced-4 offspring. mes; unc-79 ced-4 hermaphrodites contained the same reduced numbers of germ cells as mes single mutants (Figure 2). Thus, as observed with mes-3, the Mes phenotype of mes-2, mes-4, and mes-6 is not suppressed by a ced-4 mutation. We conclude that mes mutations do not lead to germ cell death via the genetically regulated programmed cell death pathway. Rather, mes germ cells probably undergo a more general, necrotic type death, perhaps as a result of physiological stress or injury.

In support of this conclusion, mes germ cells do not resemble wild-type cells dying by programmed cell death. Programmed cell death leads to nuclear condensation and formation of a button-like, refractile body (SULSTON and HORVITZ 1977 ; SULSTON et al. 1983 ). In contrast, degenerative cell deaths seen in ion channel mutants of C. elegans are characterized by enhanced membrane cycling, cytoplasmic vacuoles, and swelling of the nucleus and cell (CHALFIE and WOLINSKY 1990 ; HALL et al. 1997 ). mes germ cells display cytoplasmic vacuoles and nuclear swelling (PAULSEN et al. 1995 ; Figure 1A). Although the syncytial nature of the germline makes it difficult to directly compare the pathology of mes mutant germlines with degenerating somatic cells, we think that the mes mutant phenotype is most consistent with degeneration.

We postulated that the degeneration of mes germ nuclei might be a consequence of defects in the execution of mitosis, leading to aneuploidy or other chromosomal abnormalities. To test this idea, we constructed mes strains carrying a glp-4 mutation, which arrests germline mitosis, and investigated whether mitotic arrest would prevent the degeneration of mes nuclei. In glp-4 hermaphrodites raised at the restrictive temperature, germ nuclei are few in number (average of 12 in young adults) and appear to be arrested in mitotic prophase; they resemble some of the germ nuclei located in the distal mitotic region of a wild-type germline (BEANAN and STROME 1992 ). In mes; glp-4 hermaphrodites raised at restrictive temperature, the number of germ nuclei was the small number typical of glp-4 mutants, but the nuclei had the diffuse appearance typical of degenerating mes nuclei (Figure 3). Such defective nuclei were seen both in mes and in mes; glp-4 hermaphrodites as early as the L3 stage. We conclude that abnormalities seen in mes germ cells are probably not the result of accumulated damage from multiple rounds of faulty mitosis. Rather, mutant germ nuclei appear abnormal by L3, after they have undergone as few as two rounds of division (from two nuclei in L1s to eight nuclei in glp-4 L3s).

mes germ cells apparently adopt a germline fate, but are limited in their ability to execute that fate:
The DTC of the somatic gonad, located at the distal end of each hermaphrodite gonad arm, promotes proliferation of nearby germ cells (KIMBLE and WHITE 1981 ). When wild-type germ cells are displaced proximally by continued proliferation, they are released from the influence of the DTC, enter meiosis, and differentiate into sperm or oocytes. In mes hermaphrodites, germ cells rarely differentiate into gametes. This aspect of the phenotype could result from (1) prolonged proximity of proliferation-defective germ cells to the DTC, (2) failure in germline specification, or (3) an inability of correctly specified germ cells to execute meiosis and gametogenesis. To address which of the above possibilities is most likely, both genetic and microsurgical approaches were used to eliminate the DTC signaling pathway in different mes mutant backgrounds. This analysis was done previously with mes-3 (PAULSEN et al. 1995 ) and extended in this study to mes-2, mes-4, and mes-6.

The glp-1 gene encodes the germline receptor for the DTC signal, and failure to express functional GLP-1 causes normal germ cells to prematurely enter meiosis and differentiate into sperm (AUSTIN and KIMBLE 1987 ; CRITTENDEN et al. 1994 ). The resulting hermaphrodites have an underproliferated germline containing ~20 sperm and are sterile. To determine if mes germ cells respond normally to the absence of GLP-1, we introduced a glp-1 null allele into different mes backgrounds. Control worms showed the expected phenotypes: mes⁺; glp-1 worms lacked undifferentiated germ cells and contained ~20 sperm (AUSTIN and KIMBLE 1987 ), and mes; glp-1⁺ contained variable numbers of undifferentiated germ cells, with occasional gamete-like cells in worms bearing weaker alleles, such as mes-2(bn27) and mes-4(bn58). The experimental mes; glp-1 worms had a greatly reduced average number (at most five) of undifferentiated germ cells compared to mes; glp-1⁺ controls, indicating that mes germ cells require the DTC signal to proliferate (Figure 4A and PAULSEN et al. 1995 ). Thus, mes germ cells are normal in this aspect of response to lack of DTC signaling. However, only rare mes; glp-1 worms contained sperm (Figure 4A and PAULSEN et al. 1995 ), which were few in number and appeared defective. For example, none of the mes-2(bn27); glp-1 worms (n = 19) contained sperm, and only 5% of mes-4(bn58); glp-1 worms (n = 21) had sperm-like structures. These data indicate that when mes germ cells are deprived of the signal to undergo mitosis, they generally do not undergo meiosis and become sperm. Thus, mes germ cells are abnormal in this aspect of response to lack of DTC signaling.

Another way to assess the behavior of mes germ cells that are free from the influence of the DTC is to destroy the DTC with a laser microbeam. KIMBLE and WHITE 1981 demonstrated that laser ablation of the DTC during the mid to late L3 stage causes all germ cells in the operated gonad arm to enter meiosis and differentiate into sperm. Thus, DTC ablation phenocopies a glp-1 mutation. To confirm the above mes; glp-1 results using a different method of eliminating DTC signaling, we ablated the DTC in mes-4 and mes-6 mutant larvae. The DTC in one gonad arm of L3-stage hermaphrodites was ablated, and the other gonad arm was left unoperated to serve as a control for the behavior of the germline in the presence of the DTC. After surgery, worms were allowed to recover and mature for 48 hr, before being analyzed for the number and type of germ cells present. In the operated gonad arm of wild-type worms, results were as predicted: all germ cells became sperm. In the unoperated gonad arm of the same worm, germline development was normal. Ablations performed on mes worms resulted in the cessation of germ cell proliferation, but did not result in spermatogenesis (Figure 4B). Unoperated mes-4(bn67) gonad arms (n = 11) contained an average of 34 undifferentiated germ nuclei and no sperm. The operated gonad arms of the same worms contained an average of only three undifferentiated germ cells and no sperm. Thus, the behavior of mes-4(bn67) germ cells in a DTC-less gonad arm is like that of mes-4(bn67) germ cells in a glp-1–null background. Similar outcomes were obtained when the DTC was ablated in other mes-4 and mes-6 mutants (Figure 4B and data not shown) and in mes-3 mutants (PAULSEN et al. 1995 ): ablated arms contained fewer germ cells than unablated arms, and there was no apparent increase in meiosis or gametogenesis. These results provide further evidence that the ability of mes germ cells to proliferate is dependent on the presence of the DTC, but their general inability to differentiate is not caused by proximity to the DTC. Instead, it is likely that the inability of mutant germ cells to undergo meiosis and gametogenesis is a consequence of some intrinsic physiological defect.

The germlines of mes males are healthier than those in sibling mes hermaphrodites:
In C. elegans, the choice of sexual fate is determined by the ratio of X chromosomes to sets of autosomes (MADL and HERMAN 1979 ). Diploid embryos containing two X chromosomes (X:A ratio = 1.0) develop into hermaphrodites. Those with only a single X chromosome (X:A ratio = 0.5) develop into males. Self progeny of wild-type 2X:2A hermaphrodites consist primarily of 2X:2A hermaphrodites; 1X:2A males are generated at much lower frequency (~0.3%) as a result of meiotic nondisjunction of the X chromosome (HODGKIN et al. 1979 ). We examined the germlines of rare mes males found in our stocks and were surprised to find that they exhibit much milder germline defects than mes hermaphrodites. To fully characterize the germline phenotype of mes males, we used a mutation in him-8 (high incidence of males), which increases X chromosome nondisjunction during meiosis and results in ~37% males among the self progeny of homozygous mutant hermaphrodites (HODGKIN et al. 1979 ). We constructed mes; him-8 hermaphrodites and examined the germlines of their male and hermaphrodite offspring.

Before describing hermaphrodite/male differences in the Mes phenotype, we want to note that, as observed by CAPOWSKI et al. 1991 , the expressivity of the Mes phenotype is sensitive to genetic background. For example, we observed that the hermaphrodite offspring of mes-2(bn11); him-8 and mes-6(bn66); him-8 mothers displayed a somewhat higher average number of germ nuclei than the hermaphrodite offspring of mothers bearing the same mes allele without him-8 (Figure 5). However, the genetic background effects are much smaller than the differences observed between the XX and XO animals described below.

Several aspects of the germline phenotype are improved in mes; him-8 males compared to mes; him-8 hermaphrodites (Figure 1B and Figure 6): (1) Males have a significantly higher average number of undifferentiated germ cells, (2) germ nuclei in males are usually normal in appearance and arrangement within the gonad tube, (3) germ cells in males often undergo meiosis and differentiate into sperm, and (4) males are often fertile (data not shown).

The difference in the average number of germ nuclei between males and hermaphrodites (Figure 6) is significant for each of the mes alleles tested (see MATERIALS AND METHODS for statistical analysis). For instance, mes-2(bn27) males contained 480 ± 52 germ nuclei, while their hermaphrodite siblings contained 196 ± 129. In wild-type worms, the average steady-state number of germ cells in males and hermaphrodites is ~500 and ~1500, respectively (Figure 6 legend). Thus, mes-2(bn27) males typically contain 96% of the normal number of germ cells, while mes-2(bn27) hermaphrodites contain only 13% of the normal number.

Meiotic germ cells and gametes are observed very infrequently in mes hermaphrodites, and when present, they are few in number and abnormal in appearance. In contrast, mes males frequently contain meiotic germ cells and sperm, and the number of sperm often approaches the wild-type level (Figure 1B). The percentage of males with sperm ranged from 86% for mes-4(bn67) and mes-3(bn35) to 100% for mes-2(bn27). In the majority of these worms, sperm were too numerous to count. Because of their well-developed germlines, we routinely use mes males in matings; the success of the matings attests to the fertility of mutant males.

In addition to showing that the Mes phenotype is more highly expressed and penetrant in hermaphrodites than in males, the above analyses also reveal that not all mutant males have a normal germline; in fact, some display a severe germline phenotype. For instance, 14% of mes-3(bn35); him-8 and 4% of mes-6(bn66); him-8 males had <10 germ cells and no sperm (Figure 6). We conclude that the mes genes are essential for development of the hermaphrodite germline, but that their role is not limited to hermaphrodites.

The Mes phenotype is more severe in XX than in XO worms, regardless of sexual phenotype:
Hermaphrodites and males differ in the following: (1) sex of the germline, (2) sex of the soma, and (3) X-chromosome constitution (see WHITE 1988 ; SCHEDL 1997 ; MEYER 1997 for reviews). To test which, if any, of these differences is responsible for the differential response of the two sexes to mes mutations, we genetically altered the phenotypic sex of XX and XO mes worms. We created mes strains that carry a second mutation in one of several genes that function in sex determination (Table 1). Depending on the sex determination allele used, mutant animals were sexually transformed in the germline only or in both the germline and soma. We focused our analysis on mes-4(bn67) because, of the mes mutations characterized, this allele causes the most severe reduction in hermaphrodite germline proliferation and gametogenesis and the most dramatic hermaphrodite/male difference in Mes phenotype. We also did some experiments with a weaker allele, mes-4(bn58).

To determine whether the relative health of mes-4 male germlines is caused by some distinctly masculine quality of the germline itself, we combined mutations in mes-4 with a gain-of-function (gf) mutation fem-3. This fem-3 mutation masculinizes the germline of XX worms: instead of making oocytes, fem-3(gf) XX worms raised at a restrictive temperature make sperm throughout adulthood and are sterile (BARTON et al. 1987 ). We observed that mes-4(bn67 or bn58); fem-3(gf) XX worms raised at a restrictive temperature did not display an increase in the average number of germ nuclei (Table 1) or an improvement in germline health relative to mes-4; fem-3(gf) XX worms raised at a permissive temperature. Thus, masculinization of the germline in mes-4 hermaphrodites does not affect the Mes mutant phenotype; mes-4 germ cells exhibit a proliferation and differentiation defect that is independent of their sexual identity.

We used a mutation in fog-1 to create a reciprocal sexual transformation of the germline. This mutation causes feminization of male and hermaphrodite germlines at a restrictive temperature, causing both to make only oocytes (BARTON and KIMBLE 1990 ). When fog-1 was combined with either mes-4(bn67) or mes-4(bn58), the phenotype of mes-4 male and hermaphrodite germlines was unchanged. In particular, the germlines of mes-4; fog-1 males raised at a restrictive temperature were similar to the germlines of mes-4; fog-1 males raised at the permissive temperature in terms of both the number of undifferentiated germ cells and the presence of gametes (Table 1). The results with fog-1 confirm the results with fem-3(gf), namely that the degree of proliferation and differentiation seen in mes-4 germ cells, is independent of the sex of the germ cells.

Using loss-of-function mutations in tra-1 (HODGKIN and BRENNER 1977 ; HODGKIN 1987 ) and tra-2 (KLASS et al. 1976 ; HODGKIN and BRENNER 1977 ), we addressed whether masculinization of somatic structures would affect the Mes-4 germline phenotype. The specific mutations used cause imperfect but complementary transformations of XX worms into males. tra-1 XX males display complete masculinization of the nongonadal soma, but only one third of these animals also have a fully male somatic gonad and germline, and are able to sire progeny (SCHEDL et al. 1989 ). tra-2 XX males have a fully transformed somatic gonad and germline, but some of the nongonadal structures are incompletely masculinized (KLASS et al. 1976 ; HODGKIN and BRENNER 1977 ). Using mes-4(bn67) and each of these tra mutations, we compared the germlines of mes-4; tra XX males with mes-4; tra/+ and mes-4; +/+ XX hermaphrodites (Table 1). The average number of germ nuclei (all undifferentiated) in mes-4; tra-1 XX males (n = 39 worms) was 11 ± 6, which is very similar to the 9 ± 6 germ nuclei observed in mes-4; tra-1/+ and mes-4; +/+ XX hermaphrodite controls (n = 50 worms; Table 1). Similarly, the average number of germ nuclei in mes-4; tra-2 XX males was 14 ± 13, compared to 15 ± 14 in mes-4; tra-2/+ and mes-4; +/+ XX hermaphrodites (Table 1 and Figure 7). Thus, the severe germline phenotype of XX mes-4 worms is not rescued by masculinization of the germline or soma.

To perform the reciprocal test and determine whether feminization of the germline and soma causes XO worms to display a more severe Mes-4 phenotype, we used a gain-of-function mutation in tra-1, which acts dominantly to feminize both XX and XO worms (HODGKIN 1987 ). XX tra-1(gf)/+ worms are females, which produce normal oocytes and no sperm. These worms are self-sterile because of a lack of sperm, but they produce outcross progeny when mated. XO tra-1(gf)/+ worms display feminization of both the germline and the soma. These worms are self-sterile females, which produce variably abnormal oocytes and no sperm, and produce only a few progeny when mated. They also have stumpy tails, reflecting incomplete feminization of this somatic structure.

We wanted to compare the germlines of mes-4; tra-1(gf)/+ XO worms and mes-4; +/+ XO worms. To generate these worms, mes-4; tra-1(gf)/+ XX females were mated to mes-4; +/+ XO males. All F1s were homozygous for mes-4. The XX worms had a hermaphrodite-like spike tail, the XO tra-1+ worms had a male-like spade tail, and the XO tra-1(gf)/⁺ worms had a stumpy tail. We analyzed 45 mes-4; +/+ XO control males (Table 1 and Figure 7), which contained an average of 193 ± 127 germ nuclei. Of these 45 worms, 26 lacked sperm, 9 contained 5–30 sperm, and 10 contained >200 sperm. The percentage of sperm-producing males (42%) was lower than the percentage reported above for mes-4(bn67); him-8 males (86%), possibly because of an effect of genetic background. We analyzed 28 mes-4; tra-1(gf)/+ XO females with stumpy tails (Table 1 and Figure 7), which contained an average of 335 ± 186 germ cells. 50% of the worms contained what appeared to be endomitotic oocytes (IWASAKI et al. 1996 ); one worm contained 11 sperm. This frequency of gametogenesis is similar to the frequency observed in the mes-4; +/+ XO control males, although the type of gametogenesis (oogenesis vs. spermatogenesis) is altered by the tra-1(gf) mutation.

Two aspects of germline development were dramatically improved in mes-4; tra-1(gf)/+ XO females, compared to any mes-4 XX worms we have studied. From the standpoint of germline proliferation, the mes-4; tra-1(gf)/+ XO females contained ~23% the wild-type hermaphrodite level of germ cells, compared to a value of ~1% for mes-4; tra-1⁺ XX hermaphrodites (see Figure 6). Moreover, the mes-4; tra-1(gf)/+ XOs attempted gametogenesis at a much higher rate (54% of worms) than mes-4; tra-1⁺ XX worms (never observed). From these results, we conclude that mes-4 XO germ cells are able to undergo extensive proliferation and frequent differentiation within a hermaphrodite somatic environment and when the sexual identity of the germline is female.

The mes-4; tra-1 and mes-4; tra-2 results demonstrate that a male body and germline is not sufficient for germline health, and the mes-4; tra-1(gf) results demonstrate that maleness is not necessary. Therefore, sexual phenotype probably does not play a role in making male germlines more tolerant of mes-4 mutations than hermaphrodite germlines. The relevant factor is more likely to be the X chromosome constitution or the X:A ratio.

Our limited analysis of sexually transformed mes-2 and mes-6 worms suggests that the findings described above with mes-4 extend to the other mes genes as well. mes-2(bn11) XX worms whose somatic tissues (and usually germlines) were masculinized by tra-1 did not show an improvement in germline phenotype relative to their nontransformed siblings: mes-2; tra-1 XX males had 33 ± 32 germ cells, compared to 58 ± 55 in their sibling mes-2; tra-1/+ and +/+ XX hermaphrodites, and all worms lacked gametes. mes-6(bn66) XX worms whose somatic tissues (and usually germlines) were masculinized by tra-1 had 47 ± 49 germ cells and no gametes. Sibling mes-6; tra-1/+ and +/+ XX hermaphrodites had 50 ± 60 germ cells. Three of the latter worms (n = 43) contained a few structures that looked like oocytes. Based on these results, we tentatively conclude that, as with mes-4, the severity of the germline phenotype caused by mutations in the other mes genes does not depend on sexual phenotype but is more likely to be sensitive to X chromosome dosage or the X:A ratio.

The Mes phenotype is more severe in XXX worms than in XX and XO worms:
We described in the previous sections the less severe Mes phenotype displayed by XO mes worms compared to XX worms. To further test the sensitivity of the Mes phenotype to X chromosome dosage, we examined worms with an increased number of X chromosomes. Nondisjunction of the X chromosome results in diplo-X and nullo-X gametes and, as a result, can lead to the production of XXX offspring (dumpy hermaphrodites), as well as XO offspring (males; HODGKIN et al. 1979 ). From the broods of mes; him-8 mothers, we compared the germlines of Dpy hermaphrodites (XXX) with the germlines of their non-Dpy hermaphrodite (XX) and male (XO) siblings. Using mutations in each of the four mes genes, we observed that the average number of germ nuclei in Dpy hermaphrodites was significantly reduced below the value in non-Dpy hermaphrodites, which was reduced below the value in males (Figure 8). This observation strengthens the correlation between the dosage of X chromosomes and the severity of the Mes phenotype.

mes mutations resemble dosage compensation dpy mutations in feminizing the germline and soma of male worms whose sexual identity is ambiguous:
The sensitivity of the Mes phenotype to X chromosome dosage or X:A ratio is reminiscent of the situation with mutations in dosage compensation, the process by which levels of X chromosome gene expression are equalized in animals bearing one vs. two X chromosomes. In C. elegans, dosage compensation in the soma is implemented in diploid animals bearing two Xs: XX embryos reduce transcription from each X by half to achieve levels of X transcripts similar to those produced by the single X chromosome in XO males (MEYER and CASSON 1986 ; DELONG et al. 1987 ). Dosage compensation is an essential function in 2X:2A animals; 2X:2A embryos that fail to dosage compensate die from elevated X-linked gene expression (HODGKIN 1983 ; MEYER and CASSON 1986 ; PLENEFISCH et al. 1989 ). Conversely, 1X:2A embryos that inappropriately implement dosage compensation die from underexpression of the X chromosome (MILLER et al. 1988 ). Dosage compensation is controlled by the dosage compensation dpy genes (dpy-21, dpy-26, dpy-27, and dpy-28) and by upstream genes in the sex determination/dosage compensation pathway (Figure 9; reviewed by MEYER 1997 ). Mutations in the dosage compensation genes have a maternal effect and generally cause lethality of XX offspring, but do not affect XO worms. Rare XX survivors are abnormally short and fat, i.e., dumpy, reflecting the deleterious effects of elevated X-linked gene expression. They are also fertile, indicating that the dpy dosage compensation activities are not required for normal germline proliferation and development. Molecular analysis of dpy-26, dpy-27, and the upstream genes sdc-2 and sdc-3 has led to the following model: a complex of dosage compensation proteins assembles specifically on the two X chromosomes in hermaphrodites and reduces gene expression by partially condensing interphase chromatin (CHUANG et al. 1994 , CHUANG et al. 1996 ; LIEB et al. 1996 ; DAVIS and MEYER 1997 ).

View larger version (29K): [in this window] [in a new window]   Figure 9. —The regulatory hierarchy controlling somatic sex determination and dosage compensation in C. elegans. The diagram shows the activity states of genes that act in a regulatory cascade to control both processes in hermaphrodites (A) and in males (B). Solid letters represent a high activity state and outlined letters represent a low activity state (modified from HSU and MEYER 1994 ).

As described above for the mes genes, the mutant phenotype of the dpy genes is sensitive to X chromosome dosage: 3X:2A animals are more severely affected than 2X:2A animals, which are more severely affected than 1X:2A animals (DELONG et al. 1987 ). This similarity in mutant phenotype may indicate that the mes genes, like the dosage compensation genes, are involved in regulating the structure of chromatin and the expression of genes on the X chromosome. To explore this possibility, we tested for a further genetic similarity between the mes genes and the dpy genes. Disruption of dosage compensation has been observed to feed back on sex determination. Specifically, 2X:3A animals and XX animals carrying a mutation in the sex determination function of sdc-3, i.e., sdc-3(Tra) mutation, develop as males, whereas 2X:3A and sdc-3(Tra) XX animals in which dosage compensation has been disrupted develop as hermaphrodites (HODGKIN 1987 ; PLENEFISCH et al. 1989 ; DELONG et al. 1993 ; see DISCUSSION). To determine whether mutations in the mes genes affect sex determination, we generated and analyzed 2X:3A mes mutants and sdc-3(Tra); mes mutants.

To generate 2X:3A animals, we mated tetraploid males (2X:4A) with dpy or unc diploid hermaphrodites (2X:2A; MADL and HERMAN 1979 ). Outcross progeny, i.e., non-Dpy or non-Unc, are 2X:3A, resulting from the fusion of 1X:2A sperm with 1X:1A oocytes. In control crosses of tetraploid males to dpy or unc hermaphrodites that were mes⁺, the majority (88–99%) of the outcross progeny were male (Figure 10A). In experimental crosses of tetraploid males with mes; dpy or mes; unc hermaphrodites, the majority of outcross progeny were hermaphrodite (Figure 10A). The percentage of hermaphrodite progeny ranged from 63% for mes-6(bn64) to 93% for mes-4(bn67). These values can be compared to 55, 71, and 73% hermaphrodite progeny when the mothers carry mutations in dpy-21, dpy-27, or dpy-28, respectively (HODGKIN 1987 ; PLENEFISCH et al. 1989 ).

View larger version (18K): [in this window] [in a new window]   Figure 10. —Mutations in the mes genes can have strong feminizing effects on the soma. (A) Most 2X:3A offspring of mes+ mothers are males, but most 2X:3A offspring of mes- mothers are hermaphrodites. Diploid hermaphrodites (2X:2A), homozygous for a mutation that causes a visible phenotype, were mated to unmarked tetraploid males (2X:4A), and the sexual phenotype of their 2X:3A cross progeny was determined. (Gray bars) Offspring of mes+ mothers. (Black bars) offspring of mes mothers. At least 80 2X:3A worms were scored for each cross. (B) mes mutations suppress the masculinization of XX worms caused by a mutation that disrupts the sex determination activity of sdc-3. The mothers of the worms analyzed were sdc-3(Tra) unc-76/++, which were either wild-type or homozyous mutant for each mes gene. We examined the sexual phenotype of their Unc progeny. (Gray bar) Unc progeny of sdc-3(Tra) unc-76/++ mothers. (Black bars) Unc progeny of mes; sdc-3(Tra) unc-76/++ mothers. Between 35 and 418 worms were sc