Germ Line Sex Determination

Zygotically contributed ovo gene product is required for the survival of female germ cells in Drosophila melanogaster. Trans-allelic combinations of weak and dominant ovo mutations (ovoD) result in viable germ cells that appear to be partially transformed from female to male sexual identity. The O V O ” ~ mutation is partially suppressed by many Sex-lethal alleles that affect the soma, while those that affect only the germ line fail to interact with ovoD2. One of two loss-of-function ovo alleles is suppressed by a loss-of-function Sex-lethal allele. Because ovo mutations are germ line dependent, it is likely that ovo is suppressed by way of communication between the somatic and germ lines. A loss-of-function allele of ovo is epistatic to germ line dependent mutations in Sex-lethal. The germ line dependent sex determination mutation, sans Ji l le , and ovoD mutations show a dominant synergistic interaction resulting in partial transformation of germ line sexual identity. The ovo locus appears to be involved in germ line sex determination and is linked in some manner to sex determination in the soma. F ERTILITY depends on the proper sexual development of somatic tissues and the germ line. A great deal has been learned about the genetic control of somatic sex determination in Drosophila melanogaster (see reviews by BAKER and BELOTE 1983; CLINE 1985, 1988a; NOTHIGER and STEINMANN-ZWICKY 1985, 1987). Very little is known about the genetic control of germ line sex determination. Sex determination in the soma can be briefly summarized as follows. The primary determinant of sexual identity is the relative abundance of X-linked genes or sites (numerator elements) as compared to autosome sets (denominator elements). This relationship is commonly denoted as the X:A ratio. Flies with X:A ratios of one or greater develop as females while flies with X:A ratios of one-half or less develop as males (MORGAN and BRIDGES 19 19; BRIDGES 192 1, 1925a, b; HINTON 1955; HANNAH-ALAVA and STERN 1957; STERN 1966; SANTAMARIA 1983; CLINE 1976, 1979, 1983a, 1986, 1988b; FUYAMA 1987; TORRES and SANCHEZ 1989). Under wild-type diploid conditions the X:A ratio for female development is one (e.g., 2X:2A). In 2X:2A zygotes a number of gene products stored in the egg and the X:A ratio activate and/or maintain expression of Sex-lethal ( S x l ) (CLINE 1978, 1983a, 1984, 1986, 1988b; CHRONMILLER and CLINE 1987; OLIVER, PERRIMON and MAHOWALD 1988; STEINMANN-ZWICKY 1988). The Sxl gene product is required for the regulation of the downstream sex determination genes transformer ( t ra ) and doublesex Stanford, California 94305. Genetics 125: 535-550 (July, 1990) ’ Present address: Department of Biological Sciences, Stanford University, (dsx) and for the repression of X-linked dosage compensation (LUCCHESI and SKRIPSKY 198 1 ; SANCHEZ and NOTHIGER 1982; SKRIPSKY and LUCCHESI 1982; UENOYAMA et al. 1982; CLINE 1984; GERGEN 1987; LUCCHESI and MANNING 1987; NAGOSHI et al. 1988). The tra gene product, and the transformer-:! ( t ra-2) gene product, control the function of the dsx gene (BAKER and RIDGE 1980; BELOTE and BAKER 1982, 1983; NAGOSHI et al. 1988). The female specific dsx gene product, and the intersex ( ix) gene product, repress the expression of male specific terminal differentiation genes, resulting in female development (BAKER and RIDGE 1980; BELOTE and BAKER 1983). In chromosomal males Sxl , tra, tra-2 and ix are not required for somatic sexual identity. The absence of these functions results in the expression of male specific dsx gene product, which represses the expression of female specific terminal differentiation genes. Molecular data indicate that the gene activities of somatic sex determination genes are regulated, at least in part, by pre-mRNA splicing (reviewed by BAKER 1989; HODGKIN 1989). Germ line sex determination is not under the same genetic control as sex determination in the soma. Of the regulators of sex determination in the soma, only the X:A ratio, Sxl and one of the regulators of S x l , sansfille (snf, a.k.a. f s ( l ) G 1 6 2 1 or lir), are required autonomously in 2X:2A germ cells (GANS, AUDIT and MASSON 1975; GOLLIN and KING 1981; WIESCHAUS, AUDIT and MASSON 198 1 ; PERRIMON and GANS 1983; SCHUPBACH 1985; PERRIMON et al. 1986; SALZ, CLINE and SCHEDL 1987; STEINMANN-ZWICKY, SCHMID and NOTHIGER 1989). Chromosomally female germ cells 536 B. Oliver, D. Pauli and A. P. Mahowald (2X:2A) homozygous for mutations in tra, tra-2, dsx, ix or another maternal regulator of Sxl , daughterless (da) , can form fully functional eggs in wild-type 2X:2A female hosts, indicating that the corresponding wildtype gene products are not required in germ cells for the production of eggs (MARSH and WIESCHAUS 1978; SCHUPBACH 1982; CRONMILLER and CLINE 1987). It appears that cells of the germ line and soma must not only know their own sex, but must also communicate with each other to ensure the proper proliferation and differentiation of the germ cells. Some of the evidence for this idea comes from the analysis of the germ line phenotypes of sexually transformed flies. Because tra or tra-2 2X:2A germ cells can form eggs in a wild-type female, one might expect to see germ line structures resembling eggs in 2X:2A flies homozygous for either tra or tra-1. This is not the case. These 2X:2A “males” contain few or no germ cells and the germ cells that are seen have alternatively been reported to be arrested primary spermatocytes, or a mix between spermatocytes and oocytes [BROWN and KING 1961; R. NOTHIGER, T. WEBER and M. JONGLEZ (cited in NOTHIGER and STEINMANN-ZWICKY 1985); LEUTHOLD 1986; NOTHIGER et al. 19891. The 2X:2A “males” resulting from homozygosity for some somatic line specific alleles of Sxl also exhibit male germ line morphology (CLINE 1984). In none of these 2X:2A “males” are advanced stages of spermatogenesis seen [irrespective of the presence of a Y chromosome which bears male fertility genes (see LINDSLEY and TOKUYASU 1980)], and Sxl , tra or tra-2 2X:2A “males” fail to express at least some male germ line dependent transcripts (SCHAFER 1986; DIBENEDETTO et al. 1987; W. MATTOX, personal communication). These results suggest that S x l , tra, and tra-2 can have a somatic line dependent influence on the final number of 2X:2A germ cells in the adult gonad and the sexual identity of those cells. The 2X:2A intersexes resulting from homozygosity for dsx or ix can form early egg chambers at low frequency (SCHUPBACH 1982), suggesting that female sexual identity is maintained in at least some of the 2X:2A germ cells of dsx or i x intersexes. This may be a function of mixed somatic signals, since 2X:2A flies trans-allelic for dominant dsx mutations and loss of function dsx mutations develop as phenotypic males with germ line phenotypes similar to 2X:2A traor tra-2flies (NOTHIGER et al. 1987; B. OLIVER and R. NAGOSHI, unpublished data). All of these results suggest that 2X:2A germ cells can partially switch sexual identity (female to male) in a male soma. Further evidence that the somatic sexual environment is important for germ line development comes from mosaic studies. Somatic sex determination is cell autonomous. When a 2X:2A embryo loses an X chromosome during early nuclear divisions large patches of 1X:2A cells elaborate male structures, while the 2X:2A cells elaborate female structures (BRIDGES 1921). Because the somatic part of the gonad is formed in a different location than the germ cells, the germ cells and the somatic cells of a 1X:2A/2X:2A mosaic are often of different chromosomal sex (GEHRING, WIESCHAUS and HOLLIGER 1976). The gonads of these 1X:2A/2X:2A mosaic intersexes are often rudimentary. Further, the gonads that do contain differentiating gametes are always associated with somatic gonads of matching sexual identity (DOBZHANSKY 193 1; BROWN and KING 1962; GEHRING, WIESCHAUS and HOLLIGER 1976; SZABAD and FAJSZI 1982; B. OLIVER, unpublished data). Some atrophic ovaries contain cells of apparent male sexual identity (B. OLIVER, unpublished data), but other atrophic gonads contain no germ cells (SZABAD and FAJSZI 1982; B. OLIVER, unpublished data). It was suggested that the atrophic nature of many of the gonads of mosaics is due to a germ line/somatic line interaction resulting in the death or retarded development of germ cells that become enclosed in a gonad of the opposite sex (GEHRING, WIESCHAUS and HOLLIGER 1976). Transplantation data support the idea that the sexual identity of the soma influences the sexual identity of the germ line. 1X:2A germ cells transplanted into female hosts do not form functional eggs and can not even be found in the adult ovary (VAN DEUSEN 1976; SCHUPBACH 1985). It is a reasonable suggestion that 1X:2A germ cells die, or have such a growth disadvantage in a 2X:2A soma that they are effectively eliminated. In other experiments, using hosts that have no germ cells of their own, it has been shown that both germ cell autonomous and somatic cues are important for germ cell sexual identity (STEINMANNZWICKY, SCHMID and NOTHIGER 1989). 1X:2A germ cells transplanted into 2X:2A hosts with no endogenous germ cells appear to be arrested as early spermatocytes suggesting that the 1X:2A germ cell sexual identity is not switched in a female soma. In contrast, 2X:2A germ cells are switched to a malelike identity in a 1X:2A host. This latter observation is consistent with the somatic line dependent transformation of the germ line observed in 2X:2A flies transformed to male by Sxl , tra, tra-2 or dsx mutations. There should be genes responsible for reception of sexual signals from the soma and the ultimate sexual differentiation of germ cells. One of these genes is probably Sxl . Germ cells homozygous for various recessive alleles of Sxl differentiate cells that resemble early spermatocytes (SCHUPBACH 1985; PERRIMON et al. 1986; SALZ, CLINE and SCHEDL 1987; OLIVER, PERRIMON and MAHOWALD 1988) and constitutive expression of Sxl in 2X:2A germ cells can “override” the influence of a male soma, resulting in female germ cell differentiation (STEINMANN-ZWICKY, SCHMID and