Butterfly effect

Most of you probably know the basic story you were taught in high school biology; DNA has information-containing chunks called genes, which sit in specific places in packages called chromosomes.  Each of those genes is the instruction set for building a specific protein, and the instructions are read by creating an intermediary copy (called mRNA) of the specific gene in question, which then is transported to a structure in the cell called the ribosome, where the sequence is read and used to assemble the protein from smaller bits called amino acids.  The protein thus created -- it could be an enzyme, a structural protein, an energy carrier, a pigment, or any of dozens of other types -- goes on to do its specific job in the organism.

This pattern -- gene (DNA) to mRNA to protein -- was thought to be more or less the whole story by the people who unraveled the pattern, James Watson and Francis Crick, so in their typical self-congratulatory fashion they called it the "Central Dogma of Molecular Genetics."  (And if you know anything about their history, you'll understand why I'm calling it "typical.")  So it was a considerable shock when researchers found out that there was DNA in the genome of every species studied that didn't work this way.

It was a considerably bigger shock when it was found that the amount of DNA that did work this way was around one percent.

You read that right; between ninety-eight and ninety-nine percent of your genome is non-coding DNA.  It does not encode the instructions for building proteins.  Watson and Crick's "Central Dogma" only applies directly to less than two percent of an organism's DNA.  When this was discovered in the 1960s and 1970s, researchers (speaking of arrogance...) called it "junk DNA," following the apparent line of reasoning, "If we don't know what it does right now, it must be useless."

The whole "junk DNA" moniker never made sense to me, especially with the discovery of short tandem repeats, chunks of DNA between two and ten base pairs long that repeat over and over again (the average number of repeats is twenty-five).  Short tandem repeats are common in eukaryotic DNA, and their function is unknown.  That they do have a function -- and are not, in fact, "junk" -- is strongly supported by the fact that they're evolutionarily conserved.  Mutations happen, and if there really was no function at all for STRs, over time the pattern would get lost as mutations altered one base pair after another.  The fact that the pattern has been maintained argues that they have an important function of some kind, and mutations knock that function out and are heavily selected against.

We just don't know what it is yet.

The reason all this comes up is the discovery by a team right here in my neck of the woods, at Cornell University, that at least some of the patterns in butterfly wings are controlled by "genetic switches" in their non-coding DNA.  The team, led by Anyi Mazo-Vargas, looked at forty-six regions of non-coding DNA, and found out that a significant number of them, when disabled (a technique called gene knockout), cause huge changes in the butterfly's wing pattern.  Take, for example, the brightly-colored Heliconius butterflies:

[Image licensed under the Creative Commons Heliconius mimicry, CC BY 2.5]

Knock out a DNA sequence called WntA, and the stripes disappear; disable Optix, and the wings come out jet black.

Other than the obvious deduction -- that these non-coding sequences are acting as switches determining deposition of pigments and arrangements of the cells that generate iridescence -- not much is known about how these sequences work.  "We see that there's a very conserved group of switches that are working in different positions and are activated and driving the gene," Mazo-Vargas said.  

"We have progressively come to understand that most evolution occurs because of mutations in these non-coding regions," added Robert Reed, who co-authored the paper.  "What I hope is that this paper will be a case study that shows how people can use this combination of ATAC-seq and CRISPR [two techniques used in gene modification and knockout] to begin to interrogate these interesting regions in their own study systems, whether they work on birds or flies or worms."

So once again, as we look more closely at things, we find intricacies we never dreamed of.  What I love about research like this is that a seemingly small discovery -- that small stretches of non-coding DNA control macroscopic traits like coloration -- could have an impact on our understanding of how genetics works in general.

Truly -- a "butterfly effect."

****************************************

Reconsidering the junk

Regular readers of Skeptophilia know how much I respect science, and the women and men who have devoted their lives to increasing our understanding of how things work.  The curiosity, drive, intelligence, and creativity of scientists have provided us not only with stunning technological and medical advances, but basic knowledge about everything from the origins of life to the bizarre and counterintuitive behavior of the subatomic particles that make up all the matter in the universe.

Still, scientists are only human.  They make mistakes, misunderstand what the data mean, follow leads in the wrong direction.

Fortunately, science self-corrects.  It still baffles me when people think self-correction in science is a weakness; I call this the "Everything About This Could Be Proven Wrong Tomorrow" argument.  Why anyone would think that a system of knowledge that either couldn't detect errors, or else simply ignored them, would be preferable, is beyond me.

We had a great example of science's capacity to self-correct just this week, in a paper that came out in the journal Cell.  "Sensing Self and Foreign Circular RNAs by Intron Identity," by Y. Grace Chen, Myoungjoo V. Kim, Xingqi Chen, Pedro J. Batista, Saeko Aoyama, Jeremy E. Wilusz, Akiko Iwasaki, and Howard Y. Chang, of Stanford University, the Yale School of Medicine, and the University of Pennsylvania, sounds at first like something that would only be interesting to genetics geeks like myself.  To see why it's much more than this will take a bit of background explanation.

Our traits, and the traits of every living thing on Earth, arise through a pair of processes called transcription and translation.  DNA, as you undoubtedly know, is the master set of instructions for building everything in your body; but somehow, that information has to then direct our cells to produce brown hair or A+ blood type or resistance to malaria or any of a thousand different other features of our bodies.

The way it does that is through synthesizing proteins that then are responsible for guiding everything.  The synthesis of these proteins takes two steps.  The first, transcription, is a little like making a temporary copy (called mRNA) of the instructions from a single page of a cookbook (the DNA).  Then, a structure in the cell called the ribosome reads the copied page (the mRNA), and makes the chocolate cake or honey-glazed spare ribs or eggs Benedict -- whatever the instructions say (those finished dishes represent the proteins).

A diagram showing the process of translation [image courtesy of the Wikimedia Commons]

Our master cookbook -- the DNA in every single cell in our body -- has, according to most estimates, about 30,000 different recipes.  This gives you an idea of how genetic disorders occur -- they happen when one of the recipes has a mistake, produces too much of its final product, or doesn't get read at all.

Anyhow, back in the 1950s and 1960s, when scientists were first figuring out how all of this worked, they assumed that most of the DNA was made up of actual, readable recipes, that produced something essential for the cell.  Otherwise, why would it be there?

So it came as a bit of surprise when it was found that a significant portion of your DNA -- early estimates said it could be as much as 40% -- is "noncoding."  In other words, it's made up of recipes that don't make anything.  This noncoding DNA was derisively labeled "junk DNA" -- although why such a high proportion of our genetic material would have no function whatsoever was a considerable mystery.

I was pretty skeptical about the "junk" epithet right from the get-go.  For one thing, you'd think that stretches of DNA that had no function would eventually get scrambled by random mutations, but at least some of them have patterns (such as the tandem repeat sequences -- regions of DNA that have the same base sequence repeated over and over, and which are remarkably similar even in distantly-related species).  The fact that these patterns get preserved through millions of years of evolutionary distance indicates that changing them causes problems -- i.e., they do have some function, even if we don't know what it is.

Some "junk DNA" probably does deserve the title, of course.  We have old, damaged copies of genes floating around in our DNA, which don't ever get transcribed and simply are hangers-on from our distant ancestors.  We also have odd things called transposons, which are genes that almost act like independent life forms, copying themselves and splicing the copies elsewhere in our genomes.  (Some of those transposons are functional in switching genes on and off, but others are more like intranuclear parasites.)

Anyhow, my point is that I've long suspected that most of the noncoding DNA would turn out not to be useless after all.  And the paper by Chen et al. has just shown us that some of what seemed to be the junkiest of junk DNA -- the introns, pieces of DNA that are transcribed into mRNA but then cut out before the process of translation -- might have a function that is downright critical.

What the paper in Cell suggests is that these introns -- the leftovers bits of RNA after they're spliced out following transcription -- could have a role in the detection of "non-self" -- i.e., the basis of our immune systems.  Chen et al. write:

Circular RNAs (circRNAs) are single-stranded RNAs that are joined head to tail with largely unknown functions.  Here we show that transfection of purified in vitro generated circRNA into mammalian cells led to potent induction of innate immunity genes and confers protection against viral infection...  These results reveal innate immune sensing of circRNA and highlight introns—the predominant output of mammalian transcription—as arbiters of self-nonself identity.

Which I think is astonishing.  These chunks of RNA, which have been compared to the full-page advertisements in a magazine article that you can tear out and throw away without losing any information, might well have a role in protecting us from infection by viruses.  How exactly they do this is beyond the scope of the current study; but just the fact that this is possible will open up huge avenues for research, possibly even leading to treatments for hitherto intractable viral infections.

So what were once derisively considered useless stretches of DNA now appear to be downright critical.  All of which brings me back to my original point; that science is powerful because it has a methodology for sifting out and correcting errors or misunderstandings.  Without that, there would be no progress -- no way, in fact, for us to discern and excise the junk in our knowledge about the universe.