Current Research
The goal of my research is to understand the evolution of
developmental-genetic pathways that control key aspects of
animal morphology and behavior. Two principles guide my research:
1) A full understanding of evolution requires an understanding of
development, and vice versa; 2) Because all divergence between species
begins as genetic variation within a species, comparative studies of
development must be complemented by investigation of the evolutionary
forces shaping polymorphism in developmental genes.
An excellent system in which to study the evolution of development
is sex determination in Drosophila and related species.
Sex determination in Drosophila is arguably one of the best
characterized genetic hierarchies in development. Sex determination
is also one of the most rapidly evolving developmental systems.
For example, Sex-lethal, the key switch gene in
Drosophila sex determination, does not function sex-specifically
in fly species outside the Drosophila genus. Thus,
sex determination in flies provides a unique opportunity to study a well
characterized developmental
system with an interesting evolutionary history. I describe below my two
major projects investigating sex determination and its evolution. These
projects are complemented by computational investigations of the
evolution of regulatory gene networks, which I also describe.
Characterization of the
Drosophila melanogaster sex-determination gene intersex
and its homologues in other species
The intersex gene is required in females for proper sexual
differentiation. With colleagues from the Baker lab, I showed that
the sex-specific function of intersex is not due to
sex-specific transcription or alternative splicing, but is instead
due to the specific interaction of Intersex protein with the
female-specific protein encoded by the doublesex gene. Indeed,
it appears that Intersex provides the DNA-binding Doublesex-Female
protein with a transcriptional activation domain (see Development
129:4661, 2002 -- reprint available
here). doublesex, unlike
Sex-lethal, appears to be rather ancient, with homologues
functioning in sex-determination in nematodes and mammals. Thus, one might
predict that intersex is also conserved in function. I have identified
homologues of intersex from other fly species, from the silkmoth
Bombyx mori, and from the zebrafish, mouse, and human.
The fly homologues appear to be
conserved in function, in that they restore proper female differentiation
in Drosophila melanogaster flies that are otherwise mutant for
intersex. The moth homologue only partially restores
proper female differentiation, suggesting conserved function but
co-evolution with the Doublesex-Female protein.
The mouse homologue produces an interesting
dominant-negative effect in Drosophila, suggesting that it
misregulates target genes. I will follow up these findings with
genetic and biochemical assays of protein co-evolution, as well as
studies to determine the protein-interaction partners of Intersex
in other species and the biological processes in which it plays
a role in these species.
Functional polymorphism and divergence in the sex-determination
pathway
A fundamental question at the heart of evolutionary-developmental biology
is whether developmental-genetic systems are resistant to
change (i.e., robust to mutation), and, if so, why? In the 1940s,
Waddington coined the term canalization to describe this robustness
and postulated that natural selection would favor canalized genetic systems
because then phenotypes would depart less from their optimum. Recently,
Aviv Bergman and I showed that canalization does not require this type of
natural selection, but instead emerges as a property of complex genetic
networks that produce any phenotype (see PNAS 99:10229, 2002 --
reprint available
here). More recently we showed how
loss-of-function mutations in network genes modulate the level of
robustness, in general making more phenotypic variation
available for selection (see Nature 424:549, 2003 --
reprint available
here).
The implication for experimental studies
is that understanding the effect of mutations on developmental systems
(and therefore understanding the evolution of these systems) demands
attention to the actual network structure of the genes involved. Again,
sex determination in
Drosophila provides an outstanding model system, in
this case allowing us to test ideas about the relationship between network
structure and phenotypic divergence. In particular, I focus on the
hermaphrodite gene, which functions at two stages in Drosophila
sex-determination: early, as an activator of Sex-lethal, and late,
at the same level of the hierarchy as doublesex and intersex.
I use a conditional (temperature-sensitive) allele of hermaphrodite
to perturb sexual differentiation either early or late, and measure the
effects of such manipulations in a variety of Drosophila strains
of diverse geographic origin. Different reponses to the mutation in the
different strains imply that genetic variation exists in genes involved
in sex-determination but that this variation is normally hidden in wildtype
individuals. The presence of a large amount of hidden genetic variation
associated with the early function of hermaphrodite would suggest
that divergence in the early part of the pathway among fly species is due
to genetic drift of functionally equivalent components. On the other hand,
the absence of hidden variation in the early part of the pathway would
suggest that divergence is driven by natural selection, which sweeps clean
all variation. I have measured a large amount of hidden genetic variation
associated with the late function of hermaphrodite and will soon know
whether that is matched by variation associated with the early function.
In addition to providing a means to investigate the evolutionary forces
acting on the sex-determination pathway, this panel of strains provides a
potentially powerful tool for gene discovery. For example, the possibility
exists to correlate phenotypic responses to the hermaphrodite
mutation in different strains with genetic differences between strains,
either by quantitative trait locus (QTL) mapping or by microarray-based
gene expression assays. Thus, it may be possible to isolate genes involved
in the development of a particular sexually dimorphic tissue that would
otherwise be very difficult to find due to their pleitropic effects on
other developmental processes.
Return to
Mark's Research Page
Return to
Mark's Home Page
Go to
Baker Lab