Pharyngula

This talk should put me back in my comfort zone—developmental biology, evolution, and fish, with the stickleback story, one of the really cool model systems that have emerged to study those subjects.

What is the molecular basis of evolutionary change in nature? How many genetic changes are required to produce new traits? Which genes are used? What types of mutations? Few or many changes required?

The dream experiment would be to cross a whale and a bat and figure out what their genetic differences are. That's impossible, so they searched for other organisms with a suite of differences that were crossable...and they picked the 3-spined stickleback.

Sticklebacks are migratory between salt and freshwater, and post-glaciation, they colonized many freshwater lakes and streams. The marine phenotype is ancestral, and freshwater species have many differences...but because these differences only evolved in the past 10-20,000 years, and they are crossable by artificial insemination.

Fish are small, easy to collect, and have segregating genetic traits. They have disadvantages: no sequences, no clone libraries, no genetic markers, no linkage maps, no transgenic mehtods when first worked out. Now they have dense genetic and physical maps, a complete genome sequence, whole genome transcriptome arrays, genome wide SNP studies, and high throughput transgenics. Zoom.

Specific questions: hindlimb reduction occurs in many vertebrates. Marine sticklebacks have substantial pelvis, pelvic spines, and fins. Some of the freshwater populations have lost the hindfins in cases where predation is low, calcium is low, and there are many insect predators.

Crossing marine and freshwater with pelvic reductions identified single chromosomal region that explains 65% of the variance. They also have candidate genes: Gli3, Fgf10, Shh, Fgf8, Fgf4...most interested in ones expressed only in hindlimb: Pitx1, which maps directly to chromosome involved in reductions.

The protein coding region of Pitx1 is identical, but there is a tissue-specific loss of expression in the developing pelvic region. This is a gene of large effect.

Knocking out Pitx1 in mice yields reduced hindlimbs and death before birth. Regulatory changes are more focused and specific; cis-acting regulatory changes FTW!

Still need info. What base pairs have changed? Single or multi-step mutations? Same or different events in different populations? Are there hotspots?

Doing fine mapping of the defect: there are some populations that are dimorphic, in which the mutation is not yet fixed. They've identified a 20kb interval upstream of Pitx1 that is correlated with the differences. This sequence has been tied to a reporter gene, and it does contain a pelvic enhancer.

Can they reverse the change? Couple the 20kb sequence with a Pitx1 gene, put that in a pelvic-reduced fish, and presto, it restores a full and beautiful pelvis and pelvic spine. They think they have the right region.

Looking at different pelvic-reduced populations suggest that this pelvic control region is the target of many independent mutations. Is the Pitx1 gene predisposed to mutation? Sequence is full of repeats, might be comparable to a fragile site. This is a flexible region of DNA. Four of top ten flexibility scores in the whole stickleback genome are right in that 20kb region. The upstream coding region also exhibits other signatures of selection.

Another trait: armor plates. Marine forms are heavily armored, many freshwater species have reduced armor. Similar crosses have identified a region on one chromosome that accounts for much of the variation. Similar crosses done for skin color.

Pelvice, armor, and pigmentation mutations are all in regulatory genes. Mouse and human mutations in these same genes cause all sorts of severely deleterious effects (again, likely to be regulatory changes, not changes in the genes themselves). They are not knockout mutations in the fish, but only regulatory changes.

Same genes are involved in independent mutations in stickleback populations -- how far might this similarity extend? The genes involved in stickleback pigmentation (kitlg) are also involved in human pigmentation variants. Variations in regulatory regions of kitlg account for 20% of the pigment variation in human populations, and these regions also show signs of selection. Currently injecting constructs with fish regulatory genes into mouse embryos and getting changes in pigmentation.

They've mapped many other traits to QTLs in the fish genome, but only 3 have been dug into deeply enough -- lots of work left to do!

Interestingly, marine fish carry a haplotype for the variant alleles used in armor plating and pigmentation at low levels (0.5%-1.0%), and these are subject to selective sweeps in new freshwater populations that drive them to fixation. So we see the same alleles emerging with high frequency in different populations.

Fabulous stuff. I'm struggling to restrain myself from doing a Homer-like gurgle. Mmmm, sticklebacks.