Convergent Evolution: Regulatory Lightning Strikes Twice

Abstract

Nematodes of the genus Caenorhabditis have evolved self-fertile hermaphrodites several times. Like Caenorhabditis elegans, C. briggsae has also recruited an F-box protein to regulate the sex determination gene tra-2 during the evolution of self-fertilization.

Evolutionary biologists suffer from the lack of time machines. Until we get one, the details of how many adaptations evolved remain elusive. This is particularly true for developmental aspects, where we’d like to understand how the introduction of heritable variation and constraints imposed by developmental pathways interact with selection to produce a new phenotype. However, two favorable circumstances can offer something almost as good as a time machine: one is the existence of closely related organisms for whom lack of the adaptation in question is the primitive condition (i.e. ‘ancestor proxies’), and the other is replicated cases of the same adaptation evolving in related taxa (convergent or parallel evolution, depending on the details). For the evolution of self-fertility in the nematode genus Caenorhabditis both of these are available, and both are being used to considerable effect [1]. A new study [2] in this issue of Current Biology provides a fascinating example of how convergent evolution works at a molecular level.

In both animals and plants, sexuality generally requires mating between distinct individuals. In many species males and females produce sperm and eggs, respectively, while in others anatomically identical hermaphrodites exchange sperm in reciprocal fashion. More rarely, however, some organisms are hermaphrodites that fuse their own sperm and eggs together, allowing the animal to essentially have sex with itself. Though this extreme form of inbreeding known as ‘self-fertilization’ presents certain genetic problems, it also allows for a species to consistently reproduce at densities too low for reliably finding mates. The nematode C. elegans is the poster animal for such self-fertile hermaphroditism, which also has contributed to its success as a genetic model organism. However, this reproductive mode is rather exceptional in this group of nematodes and recently derived from an ancestor that had males and females [3,4]. Most species in the genus Caenorhabditis retain this ancestral system, and as a result have radically larger effective population sizes and more robust mating behaviors than C. elegans [5,6].

Much like C. elegans, its close relative C. briggsae evolved a strikingly similar form of self-fertile hermaphroditism [7]. In both species, animals with one X chromosome are males, while animals with two X chromosomes have a female body like their ancestors, yet produce a few hundred sperm before they switch permanently to oogenesis (Figure 1A). As a result, when males are absent, virgin hermaphrodites still produce abundant progeny. Much effort has gone into characterizing how C. elegans accomplishes this feat [8]. The genus Caenorhabditis as a whole, therefore, represents a model for studying the developmental genetic details of the repeated evolution of such a trait of major adaptive significance.

Figure 1.

Hermaphroditism in Caenorhabditis nematodes.

(A) The Caenorhabditis hermaphrodite gonad. One of the two symmetrical gonad arms of an XX animal is depicted, based on a tracing of a C. briggsae young adult. Germline stem cells are concentrated at the distal tip, where they proliferate under the influence of the distal tip cell (DTC). As cells enter meiosis, they flow away from the DTC and down the bent tube of the gonad. The first to be born differentiate as XX spermatocytes (larger blue cells) and then sperm (smaller blue cells), but before spermatogenesis is complete large oocytes become apparent distal to them (pink). No intersexual cells are seen, and after the switch to oogenesis all remaining germ cells are oogonia or oocytes. (B) Genetic models for two convergently evolved hermaphrodites. The core sex determination pathway is conserved in C. elegans and C. briggsae, as are two germline genes, fog-1 and fog-3, that act downstream of the terminal transcription factor encoded by tra-1. The orange trapezoid depicts the point of action of the germline factors that regulate XX spermatogenesis (i.e. a hypothetical sexual oscillator). FOG-2 acts with GLD-1 to repress tra-2 activity, probably through translational inhibition. The gene she-1 also acts genetically as a repressor of Cb-tra-2, although its mode of action is unknown. It is depicted as distinct from the putative C. briggsae oscillator (blue trapezoid) both because at low temperatures she-1 mutants are self-fertile and because other experiments suggest the C. briggsae oscillator probably acts on aspects of the pathway that lie downstream of the Cb-fem genes [17,19].

Although a number of other systems of repeated evolution have been explored at the developmental level, XX spermatogenesis in Caenorhabditis is unusual in two ways: first, it involves the adaptive evolution of the germline as opposed to the soma. While novel somatic features may be produced to a large extent by the evolution of transcriptional control [9], germ cell differentiation relies heavily upon post-transcriptional regulation of mRNA stability and translation [10]. Second, it involves sex determination, one of the fastest evolving aspects of animal development [11]. One thus wonders whether the same types of genetic changes known to underlie the evolution of somatic morphology will also apply to the evolution of self-fertility, and also how reproducible the changes might be in independently evolved cases. The work of Guo et al. [2] sheds important light on both of these issues, and suggests that certain classes of proteins are especially susceptible to being co-opted into germline sex determination.

Guo et al. [2] began by screening for ‘atavistic’ mutations that converted C. briggsae hermaphrodites into true females. Mutations in one of the genes, she-1 (for spermless hermaphrodite), completely feminized the germ line of XX animals while allowing nearly normal levels of spermatogenesis in XO males. In this respect, she-1 is analogous to a gene in C. elegans, called fog-2, which encodes an F-box protein essential for hermaphrodite (but not male) spermatogenesis [12,13]. fog-2 is an evolutionary ‘smoking gun’: it is the product of a species-specific gene duplication, and has acquired unique sequences that allow it to help initiate XX spermatogenesis since that duplication [14]. Guo et al. [2] used recently developed genomic resources [15,16] to positionally clone she-1. Remarkably, she-1 also encodes a species-specific F-box protein, albeit one only distantly related to fog-2. As would be expected for an F-box protein, SHE-1 interacts with a core E3 ubiquitin ligase component, SKR-1, but the proteins it presumably helps target for ubiquitin-mediated degradation remain unknown. Thus, both C. elegans and C. briggsae incorporated a recently evolved F-box protein into their germline sex determination pathways in a way that promotes XX spermatogenesis.

Guo et al. [2] next used mutations in core C. briggsae sex determination genes [17,18] to examine at which point in the sex determination pathway she-1 acts to allow XX cells to assume a male fate. This revealed that she-1 behaves as a genetic repressor of the C. briggsae ortholog of tra-2, a female-promoting gene that encodes a membrane protein interacting with several other components of the sex determination pathway. This revealed another striking parallel with fog-2, which acts in C. elegans to repress tra-2 (in conjunction with the RNA-binding protein GLD-1) [12]. However, a major difference between fog-2 and she-1 is that while the former is absolutely required for all XX spermatogenesis, even null mutations of she-1 are temperature sensitive. Guo et al. [2] suggest the intriguing possibility that the incipient C. briggsae hermaphrodite initially lacked she-1, and was only able to reliably self-fertilize its eggs at cool temperatures.

With a substantial body of work on C. briggsae germline sex determination now in hand, we can ask how this species regulates XX spermatogenesis. A useful heuristic, and potentially biological construct here is that of a ‘sexual oscillator’ that first shifts XX germ cells to the male mode, and then precisely shifts them back to the female mode more consistent with X dosage (Figure 1B). Various C. briggsae mutants [17,19] suggest that this putative oscillator acts downstream of and is repressed in males by the fem genes, which encode cytoplasmic regulators of TRA-1 (the terminal transcription factor of the global sex determination pathway), and perhaps of other factors as well. The equivalent oscillator in C. elegans is inferred to lie further upstream, at the level of tra-2 and the fem genes [8]. However, in order for the oscillator to change the sex of gametes, the activity of various pathway components must be kept within certain limits. The female fate of germ cells in the XX ancestor presumably was highly reinforced, and this canalization may have to be softened to allow an oscillator to work reliably. This may be where she-1 comes in. The data of Guo et al. [2] are consistent with C. briggsae she-1 loss-of-function mutants being feminized because they have elevated Cb-tra-2 activity. This, in turn, would prevent the putative oscillator from reaching the threshold of maleness required to produce sperm. In other words, at high temperatures the oscillator in she-1 mutants oscillates, but to no avail. That she-1 males frequently produce oocytes late in life is consistent with this model: XO germ cells should be more strongly committed to the sperm fate than those of XX hermaphrodites, and thus loss of she-1 has a weaker effect in males.

Should we think of the evolution of selfing in C. elegans and C. briggsae as the same or different? We can say that in one major respect it is the same, in that there was a striking independent recruitment of an F-box protein to regulate the same core sex determination pathway gene, tra-2, in the hermaphrodites of both lineages. But we can also say it is substantially different, because while fog-2 appears to act on or be a part of the C. elegans sexual oscillator (Figure 1B), she-1 behaves like a factor that reinforces the commitment to a hermaphrodite development, but is not part of the putative oscillator itself. Recent experiments suggest that at least part of the C. briggsae oscillator regulates germline factors other than Cb-TRA-1 [19], and similar (but unknown) TRA-1-independent germline sex regulators have long been implicated in C. elegans [20]. This indicates that we have much to learn about the patterning of sexual fates in the Caenorhabditis germ line, and even more about how it evolves.