Two revolutions have transformed the science of smell from a sleepy backwater into a dynamic and fast-moving field of inquiry. The first took place in the 1970s when gas chromatography and mass spectrometry were linked together by computer and automated. You could put a smell sample into a vial, push a few buttons, and in minutes know its main components down to the molecule. GC/MS became X-ray specs for perfumers.
For some of us, “genetics” conjures the sickly sweet smell of ether wafting through the lab during a course unit on Mendelian inheritance in fruit flies. The scent takes me back to the old Life Sciences Building at Cal, where the windows could still be opened for fresh air on a warm afternoon. Needless to say, molecular biology was in its infancy and genomics was not yet born.
Genetics still involves the occasional fruit fly, but today it’s mostly driven by biotechnology and computers. On the one side are PCR machines, automated gene sequencers, heterologous expression systems, luciferase reporter assays, and plasmid vectors; on the other are brand new statistical methods that soak up lots of computing power.
The science is hard-core and pretty daunting. This may be why so few people are aware that olfactory genomics is rewriting the evolutionary history of the sense of smell. The work is exciting and is changing the status of smell in human biology. If you want to follow along, you need a little bit of background.
As genes go, the olfactory receptor (OR) genes are relatively simple: think of them as different colored beads (amino acids) on a string. These beaded strings are found on almost every mammalian chromosome, often in clusters. Sorting these genes and understanding their function is the task of genomics. The sorting work is done by computer, but the logic is the same as laying each string of beads next to one with the most similar color sequence.
Once the strings are all laid out, we can divide them into groups based on similarity of bead color pattern (amino acid sequence). Humans have about 800 bead strings (OR genes) and rats have about 1,800. More receptors means better smell, right? Not so fast.
The bead strings sort into about 17 family groups. The working assumption is that similar bead sequences create receptors that detect similar odors. Since rats and humans have receptors in each family group, it’s likely that we perceive the smelly world in roughly the same way, though with different degrees of resolution.
The genomics jockeys also analyzed OR pseudogenes—bead strings that are missing a bead or that contain a joker bead which renders the string (gene) useless as a receptor-building blueprint. In rats, about a quarter of the OR genes are pseudogenes; in humans about half of our OR genes are nonfunctional. If an odor receptor is important for survival natural selection will keep it free of disruptive mutations. So does the high proportion of OR pseudogenes in humans mean that smell is relatively unimportant to us? That’s what some have concluded.
Pseudogenes are more frequent in primates with color vision than in species with monochromatic vision, an observation that also implies a decline in smell. (To oversimplify, the idea is that with color vision to tell you if a fruit is ripe, there is less need to sniff it.)
This demotion of the sense of smell fits with conventional wisdom from the ancient Greeks all the way to Sigmund Freud. It’s a slightly depressing thought, and one that’s hard to square with the intense interest our species shows for spices, flavors, and perfumes.
But there’s evidence that this downbeat interpretation is wrong. The OR subgenome turns out to be a hot spot of gene creation. Sure, there’s a lot of useless junk in our genetic attic, but new, fully functional receptor genes have been coming on line at quite a clip. Would we infer that cell phones are a dying technology because there are many more obsolete models than currently active ones?
The genomic analysis of olfaction has just taken a positive new turn. Instead of counting up the number of OR genes, or focusing on the proportion of dysfunctional ones, scientists at University College Dublin and Texas A&M took a new tack: classifying each species by its life history—terrestrial, aquatic, or flying.
The team looked at about 50,000 OR genes from 50 species of mammals. Using a toolkit of statistical techniques, they sorted the genes into 13 OR families based on similarity. (Several of the traditional 17 families collapse to yield the new lower estimate.) The next step was to calculate how the OR genes of each species were distributed across these families. (For example, species A might be heavy on OR families 2 and 13; species B might be tilted towards families 4, 6, and 10.)
The researchers then sorted the species by OR family proportions. Lo and behold, this corresponded almost perfectly with ecological lifestyle. In other words, species grouped together because they has more receptors in OR families 2 and 13 all turned out to be aquatic mammals. Another bunch of species sorted together because they had lots of receptors in OR family groups 1-3-7 and 5-8-9; turns out they are all bats.
As it is possible to assign taxa to their correct ecogroup based on their functional OR gene repertoire rather than phylogenetic relatedness, these results suggest that natural selection occurring through environmental niche specialization plays a large role in molding the OR gene repertoire in mammals, rather than shared evolutionary history and chance.This is cool stuff. Without knowing what odor molecules the receptors in families 2 and 13 detect, we can still infer that they are specialized for the perception of waterborne odors—all this from sorting through gene sequences.
The implications for evolutionary history are also big. The picture that emerges is of a dynamic sensory system that responds quickly to selective pressure. I see it as another blow against the depressing Greco-Freudian doctrine of nasal decline. In What the Nose Knows I speculated that the human sense of smell evolved in response to our habit of cooking and spicing food. I thought the idea was plausible but a little far out. Now I’m liking it more and more.
* * *
The new study is by Sara Hayden, Michaël Bekaert, Tess A. Crider, Stefano Mariani, William J. Murphy and Emma C. Teeling. It appeared in Genome Research. The image above is from their article; one of many dazzling graphics.