Looking at a snorting French bulldog or a prancing Pomeranian it can be hard to grasp how these pint-sized pooches could have possibly descended from wolves, which today routinely exceed 100 pounds and can take down bison.
Given the huge size difference between wolves and the littlest dogs, it’s easy to assume the genetic changes that brought about the supreme smallness of Chihuahuas and their ilk only appeared as humans started domesticating dogs, about 15,000 years ago.
Now, research published today in Current Biology suggests instead that the genetic fodder for the tiny dog breeds of today was actually lurking in the DNA of ancient wolves that lived more than 50,000 years ago, long before domestication began.
Back in 2007, Elaine Ostrander, a geneticist at the National Human Genome Research Institute and senior author of the current paper, and her colleagues discovered that a single gene is largely responsible for the size differences between dog breeds. The gene in question codes for a protein called insulin-like growth factor-1 (IGF-1).
The IGF-1 protein is a big deal when it comes to body size in mammals and dog breeds are an especially clear example. For instance, a toy poodle has much less IGF-1 protein floating around in its blood than a larger standard poodle does. On balance, the gene that codes for the IGF-1 protein controls about 15 percent of size variation among dogs while 19 other genes account for the rest. By contrast, there are roughly 10,000 DNA markers in humans that all play roles in determining how tall a person becomes, with the most influential single gene accounting for less than half a percent.
But until now researchers could never locate the mutation, or the alteration in the gene’s DNA sequence, at the IGF-1 gene region that led to less of the growth-promoting protein in smaller dogs. More than a decade later and with the help of modern genomic analysis, a post-doc in Ostrander’s lab named Jocelyn Plassais used the first Covid-19 lockdown in the summer of 2020 to take a fresh look at the IGF-1 gene region in dogs with an unorthodox approach: he started reading some of the genetic code backwards.
This genetic sleuthing hit on a particular gene that gets trasncribed into what's called long non-coding RNA, which is so named because it doesn’t code for any proteins. The team found that there were two versions or variants of the gene that codes for this strand of RNA present in domestic dogs, one of which appears to interfere with the production of the growth-regulating IGF-1 protein. Because dogs get one set of 39 chromosomes from each parent, they can end up with two copies of the same variant or one of each.
The researchers then searched for the backwards variants of the IGF-1 gene across a staggering 1,431 canid genomes. A full 1,156 of those genomes came from modern dogs (Canis familiaris), a single species which encompasses 230 established breeds, 140 indigenous and village dogs, and one dingo (which is generally considered to be a type of dog but that some argue is its own wild species). In addition to all these dogs, the study also examined the genomes of 13 species of wild canids from around the globe and 35 ancient dogs and wolves.
Ostrander says the newly identified variants or alleles “held up beautifully in dogs.” Among domestic dogs, 75 percent of those under about 30 pounds had two copies of the IGF-1 gene variant that the researchers termed the C allele and 75 percent of dogs that weighed more than around 50 pounds had two copies of the T allele. These correlations were so strong that Ostrander and her co-authors dropped the C and T monikers and started simply referring to the variants as the small and large alleles.
In wild canids, wolves usually had two copies of the large allele with a few exceptions, while non-wolf species including foxes, jackals and African hunting dogs all had two copies of the small allele. Coyotes fall into this non-wolf group as well but they show more variation since in some parts of the world, especially in the eastern United States, they sometimes interbreed with wolves. These coywolves, as they are colloquially called, tend to be bigger and often have at least one copy of the large variant of the IGF-1 gene.
Finally, to try to figure out when and where these variants emerged, the team analyzed the genomes of the oldest canines they could get their hands on. The linchpin of these results comes from the DNA of an ancient wolf that was preserved for 52,500 years in Siberian permafrost. This animal’s genome didn’t contain two copies of the large variant, instead it had one copy of the small variant alongside the large variant.
“In ancient canids the expectation was that they’d have two copies of the big allele,” says Ostrander. “But it turned out the small allele was in the population, just at a low frequency. It’s almost like nature was keeping this in her back pocket.”
Ostrander and Plassais say this suggests the small allele was present at least 54,000 years ago and may predate the large allele. The researchers estimate that the 52,500-year-old wolf with one copy of the small variant weighed about 48 pounds. Even if, further back in evolutionary time, wolves with two copies of the small allele were the norm they would not have been the size of toy poodles, because of the 85 percent of variation in canid body size accounted for by genes other than the IGF-1 gene.
“This gene was around at the time humans started to domesticate dogs, so people were able to start selecting for it,” says Ostrander. “This allowed humans to manipulate body size very quickly to breed for big dogs to guard, smaller dogs to herd and even smaller dogs to rat.”
Oxford University evolutionary biologist Greger Larson helped the team on these ancient canine genomes, and in light of these results he says it’s probably more accurate to think of the large variant of the IGF-1 gene as the mutation and the small variant as the original, ancestral state. The large variant, Larson says, may have helped ancient wolves increase their body size to adapt to colder climates as they moved into more northerly habitats.
“Not only did this team track down these two IGF-1 alleles that are super influential when it comes to body size, they figured out that the original variant was the small allele and that it predates domestication,” says Elinor Karlsson, a geneticist at the Broad Institute who was not involved in the research. “We kind of think of small dogs as being the weird ones but this shows that the larger dog and larger wolf version of this gene is actually newer or weirder.”
The relatively simple genetic architecture of dogs could also one day help us better understand how genes work in our own species. Ostrander says her lab’s goal now is to try to figure out the biological mechanism by which this tweaked piece of backwards, long non-coding RNA regulates IGF-1 and by extension body size.
Karlsson says these backwards-reading snippets of genetic code, known to researchers as antisense genes, are all over most mammalian genomes, including humans. For instance, Karlsson’s lab has studied a gene associated with bone cancer with an antisense gene next to it.
“If we can figure out the biological mechanism of how exactly this tiny change to the IGF-1 gene interferes with body size, it could give us a new way to think about what may be happening with other genes and the traits associated with them,” she says. “For genetic diseases, understanding the actual mechanism is what’s required to potentially develop a therapy to address the disease. That’s what’s at the end of the road when it comes to understanding this stuff but as this study shows it’s often more complicated than we might hope.”