Beautiful Piles of Bones: An Interview with 2017 Genetics Society of America Medal Recipient David M. Kingsley
The Genetics Society of America Medal is awarded to an individual for outstanding contributions to the field of genetics in the last 15 years. Recipients of the GSA Medal are recognized for elegant and highly-meaningful contributions to modern genetics, exemplifying the ingenuity of GSA membership.
The 2017 recipient is David M. Kingsley, whose work in mouse, sticklebacks, and humans has shifted paradigms about how vertebrates evolve. Kingsley first fell in love with genetics in graduate school, where he worked on receptor-mediated endocytosis with Monty Krieger. In his postdoctoral training, he was able to unite genetics with his first scientific love — vertebrate morphology. He joined the group of Neal Copeland and Nancy Jenkins, where he led efforts to map the classical mouse skeletal mutation short ear. Convinced that experimental genetics had a unique power to reveal the inner workings of evolution, Kingsley then established the stickleback fish as an extraordinarily productive model of quantitative trait evolution in wild species. He and his colleagues revealed many important insights, including the discoveries that, major morphological differences can map to key loci with large effects, that regulatory changes in essential developmental control genes have produced advantageous new traits, and that nature has selected the same genes over and over again to drive the stickleback’s skeletal evolution. Recently, Kingsley’s group has been using these lessons to reveal more about how our own species evolved.
An abridged version of this interview was published in the December 2017 issue of GENETICS.
What inspired you to become a scientist?
My dad died of cancer when he was 34. As a little kid I was aware that you don’t know how long you have left, and I grew up wanting to make sure I spent the time I have doing something interesting and important. I thought that tackling age-old mysteries about life’s origin and mechanisms was a good way to spend my life.
What did you learn from your first mentors?
I was a kid who loved dinosaurs and skeletons. That interest was nurtured by a great high school teacher, Jack Koch at Roosevelt High School in Des Moines, Iowa. I dedicated my PhD thesis to him. In his advanced biology class we memorized the names of every bone and muscle in the cat and human skeleton. A lot of people hated it, but I loved it because you could see so much about the function and lifestyle of the organisms from the size and shapes and patterns of bones. I’m lucky because I still work on skeletal anatomy and evolution!
In graduate school, I fell in love with the power of genetics. I had a set of teachers at MIT, including David Botstein and Monty Krieger who helped me learn that with genetics you didn’t have to assume anything about the answer. You didn’t have guess you were looking for a particular type of molecule or anything like that. Genetics was an algorithm that would take you to the key components controlling a biological system no matter what they were. I saw how genetics had the power to dissect old, hard problems like cell cycle and development, which had been mysteries when I first came across them in biology class.
Why did you choose to work on the short ear gene?
As a postdoc, I got to bring together my love of genetics with my love of vertebrate morphology — I went to a mouse genetics lab where they were among the first to walk down chromosomes and identify the molecular basis of classic mouse mutations. In graduate school, I had heard a great seminar from David Hogness from Stanford, who was carrying out some of the first chromosome walks to the homeotic genes in Drosophila. Here was someone studying one of the most interesting morphological problems you could imagine: how to turn one body part into another. He was turning morphology into genes and DNA and sequence and development, and I thought that was electrifying. I could see that mouse would go through the same revolution that had come to fly.
Vertebrate genetics takes a long time, so you should pick your problem carefully. I didn’t want to pick something that was better studied in bacteria, yeast, or powerful invertebrate systems. The skeleton was perfect; it’s the defining feature of vertebrates. It also plays such an important role in animals’ external appearance that many classic mutants had already been picked up in simple morphological screens.
Near the end of grad school, I took out from the library “Genetic variants and strains of the laboratory mouse” and read the whole book—one mutant after another. We decided to go for the short ear gene, which had been worked on for decades by the person who put that wonderful book together—Margaret Green from the Jackson Labs. She was both a very perceptive scientist and a great editor and collator. So, I felt like I was dipping into one of her favorite mutations, but there were also practical reasons to choose short ear. After World War II there had been a lot of interest in the effects of radiation on the mammalian germline, and there were two big mouse forward mutation experiments in the UK and US. They both used a test strain carrying seven homozygous recessive mutations with visible phenotypes. These were six pigment mutations and short-ear.
Millions of wild-type mice were mutagenized and crossed with the test strain to measure the rate of recovering new alleles at any of the seven loci. As a result, there were lots of newly induced mutations, including a whole set of deficiency chromosomes that took out both short ear and one of the closely linked pigmentation loci. We essentially had the equivalent of a Drosophila genetics playground for this particular region of the mouse genome! We would be able to orient ourselves using the same kind of deletion breakpoints that Hogness had been using in flies. And my postdoc advisors Nancy Jenkins and Neal Copeland had already found a retroviral insertion that caused the closely-linked dilute coat phenotype, so we even had a good entry point that was within a millimorgan of the short ear gene. That was one of the reasons why I chose short ear out of the 150 or so classic skeletal mutations.
What did you learn from the short ear project?
It took about five years to do the chromosome walk in the region, and I was already an assistant professor by the time we eventually isolated the gene for this classic skeletal trait. But it was incredibly gratifying. The gene controlling skeletal morphology encoded a secreted signal already named a “bone morphogenetic protein” (BMP).
It had been named by biochemists who found that if you took an adult bone and ground it into powder and injected it under the skin of an animal, there was some magic ingredient that could generate a brand new bone at the site of implantation. And if you put the implant in the shape of a circle, for example, it would come out as a circular bone, so you could even see that the pattern in which the signal was expressed controlled something about the rough shape and morphology of the bone that resulted.
The short ear mice provided the first genetic evidence that BMPs were the endogenous signals that vertebrates were using to set the form and pattern of skeletal structures. If you had a mutation in one of the BMPs you very selectively removed the aspect of skeletal morphology controlled by that particular member of the BMP family. The short ear deficiency strains turned out to be important because they included 29 alleles at the short ear locus, of which half a dozen were regulatory mutations disrupting the flanking DNA. These later helped us to identify a whole series of modular, remarkably specific enhancers controlling different aspects of skeletal morphology. We think of them as anatomy elements because they might control expression for example in just the ribs, and maybe only in a 90-degree sector on the outside of the ribs. There would be a different controller for the inside of the ribs, allowing you to tune the overall shape. For someone originally interested in those beautiful piles of bones, to be able to break down their shapes into the expression patterns of secreted signaling molecules was an incredibly satisfying answer.
Why did you choose sticklebacks?
If you can find a way to turn old biological problems into genetics problems, then you can often find the answers to even intractable questions. A brave postdoc Katie Peichel and I spent a really fun summer in 1998 figuring out how to turn classic evolutionary questions into a genetics problem. We wanted to identify the number and type of genes and mutations that control species differences in nature. The trick was to figure out some way to cross different species, which sounds paradoxical because one definition of species is that they are reproductively isolated. The loophole is that reproductive isolation can occur through either postzygotic or prezygotic mechanisms. Postzygotic mechanisms include inviability and sterility, which are obviously hard to overcome. However prezygotic isolating mechanisms are things like behavioral or mechanical incompatibilities in mating, which can be overcome using artificial fertilization in the laboratory.
We went around talking to biologists, reading all kinds of books, looking for very young species with recently evolved dramatic skeletal differences that could still be crossed in the laboratory. We looked at wild mice and birds, but the thing that was attractive about fish was the clutch sizes tended to be very large. With a bird system, nests might have a couple of eggs, while a fish nest would have hundreds of thousands. For using a genetic approach, especially for mapping complex traits in the wild, the bigger the family size the better.
Somewhere in the middle of that summer, I found a great book chapter by Mike Bell of Stonybrook University talking about all the cool skeletal traits that had evolved in sticklebacks after the end of the last Ice Age. There was a remarkable previous literature on stickleback morphology, ecology, and behavior in new freshwater streams and lakes. And new forms had evolved not just once but thousands of times. That was because their main ancestors were migratory like Salmon and would come from the ocean into coastal areas to breed every spring. So, when the glaciers melted and lots of new lakes and streams formed, it generated all these brand new, empty environments that were colonized by sticklebacks. It was like nature had set off a replicate series of evolution experiments 10,000 years ago, producing new forms over and over again. That was beautiful to us because not only could we figure out how evolution worked in a particular lake or stream, but the system as a whole would make it possible to tell whether the mechanisms used in evolution have any repeatability to them. Is it going to be different every time? Or are there rules and principles that underlie the way organisms adapt to new conditions.
What did you learn about repeatability of evolution?
I had a debate with a fellow faculty member when I started the project because he thought the project was not worth doing. Firstly, because evolution is complicated, and if it’s controlled by lots of genes with tiny effects you’ll never find anything. But his killer argument was: even if you could do it, he wouldn’t care. And his reason was related to repeatability. He figured we would knock ourselves out trying to figure out what happened in one lake and all we would find would be historical minutiae that accumulated in that particular location, and that if you then studied a second place you’d get a different answer and then a third would give you a different answer again. It would just turn out to be postage stamp collecting, and there wouldn’t be any generality.
At the time, we didn’t have evidence one way or another. But my best reply was: how do you know? That was the great thing about genetics – it would tell you the answer no matter what the answer is. We could have learned that all the traits are controlled by tiny effects that are almost unmappable. And we could have gotten the answer that all those little tiny effects are distributed across the genome in a way that is just due to history. But that’s not what the genetics showed.
We started crossing these fish with huge skeletal differences. And by huge, I mean thirty-fold differences in the number of plates along the anterior-posterior body axis, or complete presence or absence of an entire fin, or doubling the number of teeth, or black fish vs. white fish, the kind of dramatic changes you would normally see between different genera of wild species. And we found that while none of these evolutionary differences were simple Mendelian traits, they typically had genetic architectures with one or two chromosome regions showing very large effects, perhaps explaining up to three-quarters of the variation, along with a handful of other modifier regions controlling five to ten percent of the variance. So, the genetics was manageable.
And if you compared the results from crosses done in different lakes, it tuned out the very same chromosome regions were being used over and over again in different populations. So even before we identified the genes, we knew this was going to be both interesting and doable.
We’ve subsequently taken lots of traits down to genes and molecules. We’ve found that key signals and transcription factors that developmental biologists have been studying for years turn out to be the same molecules that nature is using to redesign anatomical features. And we’re finding the reuse isn’t just from lake to lake, it’s from organism to organism. For example, although we didn’t set out to test any particular candidate genes, the genetic data showed us that some of those stickleback skeletal traits are controlled by the same kinds of bone morphogenetic proteins that we found in mouse.
How does the stickleback work all connect with your studies of human evolution?
We’re interested in why particular genes are reused throughout evolution, and we’re also interested in applying the patterns we’ve found in sticklebacks to the evolution of ourselves. We’ve found that classic traits in people, like blond hair color, or height, are evolving in humans using the same types of key control genes and regulatory mutations we have found in fish. And unlike rare genetic diseases, there are derived alleles at these human loci where a large fraction of the population carry the selected version. So rather than studying diseases that affect 1 in 100,000 people, it’s been really interesting to study variants that, because they have been subject to selection, are now present in billions of people. In some cases, the selected alleles may actually increase susceptibility to late-onset diseases like cancer or arthritis. It’s not a huge effect, maybe 1.3 to 1.8-fold. But when an allele slightly increases risk of a disease and is carried by a few billion people through selection, then suddenly you find an awful lot of the burden of a common human disease is controlled by our own evolutionary history.
We’re now going back and forth between humans and the patterns we see in fish. We thought it might take us 50 years to get enough examples to pull out general principles, but it turned out to be much faster than that. We now have a whole bunch of genomic regions—maybe 200—that have been repeatedly selected in stickleback. We’ve been able to answer how often evolution uses coding versus regulatory genes. That question was debated a long time. However, we can now say empirically the answer is both, but 85% of the time it’s regulatory and 15% coding. When I say there’s things we learned from fish that we apply to other organisms, we’re already applying things like that 85 percent rule in our human studies. If regulatory changes are by far the most common way to preserve viability and fitness when sticklebacks are evolving under a whole range of fitness constraints. then I think the things that make us human are likely to also be regulatory. So, we can prioritize our human work using the rules we learned from repeated evolution in stickleback.
What’s the best advice you ever received?
Genetics can be used to study anything.
What advice would you give to younger scientists?
Genetics can be used to study anything! I fell in love with genetics watching it be used by people who loved it. It’s such an honor to receive this award because I feel like I’m continuing that tradition— especially since it has previously been given to many of my own teachers and heroes in the field. I hope my students will also be convinced of the power of genetics and will use it to study their own favorite problems as well.