One of the unsolved problems in understanding how genes control sex determination in C. elegans is that of how spermatogenesis, a male function, occurs in the hermaphrodite, which is otherwise female. The current model for sex determination can only explain how male vs. female development occurs. As described in the last Newsletter, I have been working on several dominant
tra-2 mutations that may bear on this problem. These mutations eliminate spermatogenesis in XX animals, to produce females instead of hermaphrodites, but do not generally show effects on XO males. In the XX animal,
tra-2 is thought to act with
tra-3 to prevent the function of the fem genes. In the XO animals,
tra-2 appears to be turned off by the action of
her-1, thereby allowing the fem genes to promote male development and spermatogenesis. Since the fem genes are also needed for spermatogenesis in the XX animal, there must be a way of allowing them to promote a brief period of spermatogenesis without interfering with female somatic development or eventual oogenesis. [See Figure 1] The
tra-2(dom) mutations indicate that this transient fem expression is mediated via the modulation of
tra-2 activity. That is, that although
tra-2(+) is on in the XX animal, it is turned off transiently to allow the fems to promote spermatogenesis. In
tra-2(dom) animals,
tra-2 activity is apparently no longer modulated, i.e. it is constitutive, because
tra-2(dom) XX animals are females. However, XO
tra-2(dom) animals are usually normal males, therefore
tra-2(dom) is still turned off by
her-1 in the XO animal. In sum, in addition to the negative control of
tra-2 activity by
her-1 in the XO animal, tra- 2 is also modulated in the XX animal to allow hermaphrodite spermatogenesis to occur. It is possible that genes that specifically control this modulation exist (indicated by the ? in model). Two
tra-2(dom) alleles,
tra-2(
e1940) and
tra-2(
e2020), have been characterized in greater detail. Both have been reverted to yield several intragenic recessive masculinizing mutations that fail to complement
tra-2(
e1095) (a putative null). This indicates that they are
tra-2 alleles. Although both mutations appear to be constitutive, their penetrance differs:
tra-2(
e1940) and
tra-2(
e1940)/+ XX animals are sometimes slightly self-fertile hermaphrodites instead of females, while
tra-2(
e2020) and
tra-2(
e2020)/+ XX animals are always female. Further,
tra-2 activity appears to differ in the two alleles;
tra-2(
e1940) activity is possibly lower than wildtype (hypomorphic) because: 1)
tra-2(
e1940)/(
e1095) XX animals are self-fertile intersexes in contrast to
tra-2(+/e1095) XX animals, which are essentially wildtype