About the “Under the Radar” series: Some scientific concepts come up again and again in interviews with scientists but never find their way into newspaper headlines. Each post in this series follows one of those biology “bogeys” that fly under journalism’s radar through 3 different mini-stories.
Story #1: Scientists splice up a CRISPR chicken…and find an evolutionary shortcut
Birds’ brains have all of the tools to make mammal-like neurons, according to a study in Science from August. And, incredibly, the researchers behind the study only had to tinker with one gene that changes how chicken cells edit their RNA to unlock several seemingly unrelated mammal neuron traits in chicken neural precursor cells.
It was as if the chicken cells instantly acquired a whole bunch of mutations at once, instead of just one.
Researchers think that this gene editing process– aka “alternative splicing”–may explain why some traits seem to have evolved at such high speeds.
“This is a process that has diverged very rapidly during evolution to produce different versions of proteins,” University of Toronto geneticist Ben Blencowe explained in a phone interview.
500 million years is a long time to evolve, but it’s still hard to account for all of the diversity in vertebrates based on variation in DNA base pairs alone.
The key to animal diversity lies in an aspect of biology that your high school biology class kinda sorta covered, but lots of people forget all the steps after they’re done cramming for the test.
DNA cannot control an organism’s traits all by itself. For a gene to become a trait, the cell has to send an RNA copy of the gene to a tiny molecular factory called a ribosome. Ribosomes’ job is to build molecular machines called proteins. Proteins are the ones that run around the cell, doing all the little biochemical tasks that keep you alive, and they’re the ones that determine the final trait– usually (more on that in a later post).
Before an RNA leaves the nucleus, it has to get past little RNA editors called spliceosomes which cut unneeded base pairs out. Chickens’ spliceosomes are pretty similar to mammals’, but not identical. So when the researchers replaced one of the chicken cells’ RNA-editing genes with the mammalian version, they ended up with chicken “neural precursor” cells that build mammals’ neural proteins.
Blenclowe’s team concluded that switching up splicing patterns is big part of how animals keep up with rapidly changing evolutionary pressures, and his team isn’t the only one saying so. An MIT team lead by Christopher Burge came to the same conclusion based on a separate set of evidence. Burge’s team also pointed out that many of the genes that have alternative edits are also genes that play big roles in cancer.
Meet the Bogey: Alternative Splicing
Thanks to alternative splicing, most genes can code for more than one trait. That ability comes in handy when big, multicellular bodies need to make different kinds of cells. If one version of Protein A works best for neurons and another slightly different version of Protein A is best for muscle cells, then it’s helpful to have a single gene that can code for both. All you need is a tiny chemical cue that shows up in neurons but not muscles, and Presto! When the RNA-building and editing machinery sees that chemical, it’ll switch to making the neuron version.
If all of this is giving you a headache, trust me: you are not alone.
(The way I remember it is with a baking analogy: Say that you and your friend share a cookbook, but when you make chocolate chip cookies, you like to swap out the chocolate chips with M&Ms. If one of your friends is a stickler for following the recipe exactly, their chocolate chip cookies will turn out pretty similar to your M&M cookies but there will be slight differences.
If a third friend also uses the cookie recipe but forgets to add the salt, altogether you’ll end up with three sets of cookies, based on the same recipe, which are very similar in terms of overall chemical composition but may taste or look slightly different.
Same principle with alternative edits of the same gene: The differences in the final protein or trait are usually pretty small, but they can change behavior or traits. Also, while some alternative splicing is accidental– like leaving out the salt–many splicing variants are actually deliberate and beneficial– like using M&Ms instead of chocolate chips.)
For a long time, geneticists have accounted for the diversity of proteins by arguing that the vast swaths of non-coding DNA– sometimes labeled the “junk DNA”– are controlling gene expression, And while that is largely true– lots of non-coding DNA contains cues that tell the RNA-builders which genes to copy– biologists didn’t realize how common alternative splicing is in the active protein-coding genes.
At the turn of the millennium, most geneticists thought that only 6% of genes had useful alternate edits– aka “splicing isoforms“– but as it became easier and easier to sequence genomes and track RNAs, the number of genes known to have splicing isoforms skyrocketed. “Now we know that it’s 90% or more [of all genes],” UCSD computer scientist Christian Barrett told me during a phone interview.
Any process that occurs that frequently in living cells has to be useful, but figuring out how to crunch the data we have on splicing isoforms and convert it into medically useful information is proving to be a daunting task…
Story #2: Do cancers have unique splicing “signatures”?
The aforementioned Christian Barrett works for pediatric geneticist Kelly Frazer of UCSD. They’ve been working on a developing computer programs that can sort through the multitudinous RNAs in cancer cells and identify particular splicing variants that only occur in ovarian cancer. In a paper published in May, they announced that they had found several RNA splicing isoforms that frequently appeared in ovarian cancer cells but almost never showed up in healthy ovary tissue.
The human genome has about about 20,000 genes, but those 20,000 genes can make hundreds of thousands of different proteins. There are about 300,000 known (or at least, genomics’ most reliable computer programs say there are 300,000) protein isoforms in human cells. In other words, there’s evidence that alternative splicing produces an average of 15 different proteins for every individual gene.
Developing a computer program that could glean that useful cancer-identifying information from a sea of hundreds of thousands of RNA transcripts took Frazer and Barrett’s team about five years. And in the process, they realized that the previous estimate of total human splicing isoform products was actually probably too low. Based on the data they were seeing in ovarian cells, the total number could be closer to 400,000 or even 600,000.
“I think at lot of people would have just given up give how complex it was to design this experiment,” Frazer said.
Still, she’s optimistic that using computers to sift through RNA transcripts could lead to more accurate diagnoses. Her team is currently doing an exploratory trial to see whether the RNAs their software highlighted are present in cells from women who are currently receiving treatment for ovarian cancer.
If all goes well, they may be able to develop a simple test where doctors can take a few cells from a pap smear sample and check to see if there are any cancer-correlated RNAs in them. A test like that wouldn’t be a silver bullet, but it would probably help doctors catch ovarian cancers earlier.
But splicing isoforms’ biological importance isn’t limited to disease settings like cancer…
Story #3: Can’t sleep? Alternative splicing might be partly to blame.
Foster and Pierson are neuroscientists, not geneticists, so they measure the protein’s effects by recording how often melanopsin-expressing neurons fire and how long each of those neuron firings lasts. Most neuron-controlling proteins cause neurons to fire in one, more-or-less predictable pattern. But when they tested melanopsin, they consistently saw two distinct neuron firing patterns.
“We were getting two messages and two proteins and thought ‘What the HELL is going on here?!'” said Foster.
Geneticists know enough about splicing to predict which portions of genes are most likely to be edited out, so when Foster and Pierson asked a geneticist colleague to take a look at melanopsin’s DNA template, they were able to ID a section in the melanopsin gene that looked highly spliceable. That meant that most lab mice probably had both a “long-tail” version of melanopsin and a “short-tail” version in their retinal cells.
But when the researchers made mutant mice that could only express one of the two melanopsin isoforms, they found an even bigger surprise: The two melanopsins control separate behaviors.
The mice without short-tail melanopsin couldn’t shrink their pupils in response to bright light. However, they still responded to light the way most respectable nocturnal lab mice do: by slowing down and acting sleepy.
In contrast, the mice without the long-tail melanopsin could adjust their pupils to the light, but the light didn’t make them sleepy. Instead, they more or less behaved like tiny rodent insomniacs. Foster compared their lack of circadian rhythm to a case of neverending jetlag.
So the group’s next question is: What do melanopsins do in humans?
The human version of the melanopsin gene is pretty similar to the mouse one in terms of DNA base pairs, and it has the potential to be spliced into long-tail and short-tail versions. It seems plausible that human long-tailed melanopsin could help us adjust our pupils, but since we’re active during the day, all bets are off when it comes to what the human short-tailed melanopsin might do.
Humans actually exhibit the inverse of the mouse response to bright light; we get sleepy and sluggish when lights go dim. Foster says that if you were one of those people who drifted off to sleep during dimly lit lectures in college, “Not all of that would be the fault of your lecturer.”
So even though it’s kind of tempting to speculate that a splicing isoform controlling our response to dimly lit classrooms might be partly responsible for why so many people forget about RNA after they’re out of school, at this point, we really don’t know.
There are obvious ethical problems with knocking out protein isoforms in human subjects, especially when there’s reason to think doing so might put them at risk of insomnia.
Foster and Pierson are, however, on the lookout for humans who may have natural mutations affecting their melanopsin splices. However, only time will tell if they’re able to find someone with an identifiable melanopsin mutation.
Still, their study is one of the first to show a single splicing event creating two proteins that control two different behaviors. It probably won’t be the last…
Cells can build several different things from just one gene. This makes sorting out which gene causes which trait a lot more complicated.