I joined Dr. Craig Albertson’s lab as a graduate student in 2009, where I quickly became fascinated by these cute cichlid fishes. They’re colorful, they breed their young in the mouth, and some of them have funny looking faces like this blue mbuna (Labeotropheus fuelleborni):
My research started on the genetic control of bone development in these cichlids, trying to understand the development and evolution of their facial shape. We soon discovered that the ptch1 gene, a member of the Hedgehog signaling pathway, plays a major role in shaping multiple bones in the cichlid head. In brief, some cichlid species, such as the blue mbuna above, have one version of the ptch1 gene that leads to a facial skeleton suitable for scraping algae off the surface of rocks; some other cichlid species have a different version of the ptch1 gene which leads to a slightly different skeletal shape that make them good at capturing evasive prey via suction feeding. These findings add up to a fast-growing body of knowledge about the genotype-phenotype map, which has been of great interest to biologists for decades.
While I’m proud of my contribution as a young scientist, one thought often haunts me: although I call it a “major role”, the genetic difference in ptch1 only accounts for 10-20% of the variation in bone shape. The vast majority of the variation remains to be explained, what’s going on there?
From a geneticist’s point of view, this is exactly what you would expect to find in a trait as complex as the shape of facial skeleton: there’re probably many genes working together to produce the final phenotype. Some genes, like ptch1, can have a relatively large effect and will pop out in our analysis. But most of genes will have such a small contribution that could be easily swamped by all kinds of background noise and therefore, would be very difficult to detect. Considering the size and life history of cichlids, it’s logistically impossible for me to trace down those genes.
How about environmental factors? Genetics is certainly not the only player, I should take a detour.
I was really intrigued by a plasticity experiment in the lab, where we fed the cichlids with different food and induced skeletal changes in the face. And I started wondering, is there any other behavior that can also change their bones?
Then, the gaping behavior of cichlid larvae quickly caught my eyes.
6 day old cichlid larvae gaping
The cichlids we study are mouth brooders. Mom usually keeps all her eggs in the mouth for several weeks. But in order to study their development, we will extract the embryos from the female’s mouth and raise them in a flask. This is one of the first skills I learned when I joined the lab, and I noticed this gaping behavior right away – well, dozens of cute little baby fish gaping their mouths is kind of hard to miss. I didn’t pay much attention to this behavior at that time. And I think probably most people would have thought the same thing when they first see the video above – these larvae are just doing their normal fishy thing, they’re just breathing. It looks cute, but that’s pretty much it.
However, when I looked at this behavior with the new perspective, I soon realized its bizarreness: they were gaping really fast. Some of them would gape more than 250 times in 1 minute, so fast that I could barely count it in real time. And I had to count in Chinese, English numbers were just too long for me. That felt like a lot of gaping, were they really just breathing? A quick literature search said no: fish larvae at such a young age rely mostly on their skin for breathing. I was excited to find this out. Initially, I thought maybe this larval breathing behavior could influence bone development as a side effect. Now I got a new idea: perhaps it’s not about breathing at all. They might actually be working out their mouths to stimulate bone development.
As a graduate student, the first thing you need to do when you have an idea like this is to convince your adviser that you’re not crazy, that it’s a legit scientific hypothesis. So, I started to collect some preliminary data, which turned out to be encouraging: I found that the larvae of the blue mbuna, who has a longer jaw bone called the retroarticular process (RA), gaped at a higher frequency than another species with short RA. This is exactly what I expected to see, more gapes more bones.
At this point, both Craig and I became very interested in this hypothesis, and started to brainstorm ways to test it. We thought about all kinds of ways to manipulate the gaping behavior, from varying environmental factors like oxygen and temperature, to chemical treatment like caffeine and tricaine (commonly used for anesthesia in fish), to more goofy ideas like botox and gag. Most of them were rejected either because they would cause an overall impact on development, or simply not feasible in our fish. In the end, I decided to take a rather adventurous route, that was to surgically cut the ligament attached to the RA. We liked this approach because it’s a local and targeted experiment: the surgery would only attenuate the mechanical stimuli being applied directly to the RA without affecting the other bones and muscles that participates in the gaping behavior. The surgery worked beautifully: when I cut the ligament in the larvae of the fast gaping species, they developed a shorter RA.
The surgery was sort of a loss-of-function experiment, and we also did a gain-of-function experiment that originated from what I call a happy accident in the lab. One day I ran out of standard containers for the larvae when I was monitoring their gaping behavior, so I had to substitute with small beakers. Then I found that when restricted to a smaller environment, the larvae will gape at a higher frequency. Although it ruined my gape counts for that brood, I took advantage of this phenomenon and tried to induce gaping behavior in the larvae of slow-gaping species with smaller container. Just as I expected, they developed longer RAs.
These two experiments suggested that the seemingly trivial gaping behavior of cichlid larvae affected the development of their RA development, and the effect size was, interestingly, on par with what we found with genetic difference in the ptch1 gene. But what’s more interesting, is that we found larvae that gaped more is expressing more ptch1.
Taken together with our previous studies, we think the Hedgehog signaling pathway plays a dual role in shaping the facial skeleton. On one hand, genetic differences in the ptch1 gene leads to differential bone growth in the facial skeleton; on the other hand, mechanical stimuli also induce bone growth via Hedgehog. At this point, we’re curious about how these mechanisms evolved. It will be super cool if the epigenetic (or plastic, whichever way you prefer to call it) route came in place first, and genetic mutations came later to fix the difference – genes as followers.
And for me personally, I really want to learn the larval gaping behavior in nature. My observation of the gaping is done in beakers and Petri dishes after all. And I just can’t stop thinking what the larvae are doing in their mother’s mouth, which is a much smaller environment than my restriction experiment. Do larvae gape even faster? Does the mom modulate their gaping? Or perhaps they don’t gape at all? These questions won’t help us cure cancer. They’re not major challenges with great impact in academia or industry. I’m just curious, simple as that, which is exactly why I love doing science.
Baby fish working out: an epigenetic source of adaptive variation in the cichlid jaw. Yinan Hu, R. Craig Albertson. 2017. Proceedings of the Royal Society B: Biological Sciences.