In March, I wrote a post on some tantalizing new findings about the secrets of human evolution lurking in our genome. In brief, researchers at the University of Pennsylvania studied a gene called MYH16 that helps build jaw muscles in primates. In our own lineage, the gene has mutated and is no longer active in jaw muscles. Perhaps not coincidentally, we have much smaller, weaker jaws than other apes. The researchers estimated that the gene shut down around 2.4 million years ago--right around the time when hominid brains began to expand. They suggested that shrinking jaw muscles opened up room in the hominid head for a larger brain.
It's a cool hypothesis, but it may not hold up. Scientists at Arizona State University have followed up on the initial study by anlayzing much larger pieces of MYH16, both in humans and in other species. All told, they studied 25 times more DNA from the gene. In a paper in press at Molecular Biology and Evolution, they report finding a significantly different date for when the gene mutated. Instead of 2.4 million years ago, they get a much older date: 5.3 million years ago.
If that's true, then you can forget any significant link between the evolution of MYH16 and brain evolution. If the Arizona State team is right, the two events are separated by three million years. What's more, the jaws of hominids also remained relatively large after the mutation of MYH16.
The Arizona State researchers do point out an intriguing clue that may eventually lead to a solution to this paradox. The mutation that the Penn team originally argued that the MYH16 gene became useless when a section of DNA in the middle of its sequence was accidentally deleted. Often, when this sort of deletion takes place, DNA-copying enzymes come to a screeching halt at the site of the mutation. With the gene only partly copied, it cannot be turned into a protein. But the Arizona State researchers found signs that the gene did not shut down entirely 5.3 million years ago. The DNA "downstream" from the mutation--in other words, beyond the point where the enzymes stopped copying the gene--has picked up mutations in a pattern that shows no sign of natural selection at work. That's what you'd expect from DNA that doesn't make a gene, since any change will have no effect for good or bad on its owner. But the upstream DNA--the part of the gene that could still be copied--told a different story. It showed signs of having undergone selection. So perhaps the mutation that occurred 5.3 million years ago didn't actually kill the gene, but just amputated it. What the surviving portion of MYH16 did (or still does) remains unknown.
I would wager that this new paper will unfortunately not attract much press. When scientists first come up with an attention-grabbing hypothesis, they're more likely to get a paper accepted to a high-profile journal, and more likely still to get written up by science writers like me. But follow-up work often ends up in the shadows.
That's a shame, because science is actually not made up of single studies that suddenly overturn everything that came before. It's more of a dialectic, as different groups of scientists search for new evidence in order to put hypotheses to new tests. Some hypotheses--such as the idea that chimpanzees are our closest living relatives--have become stronger over time. Others fall away. It would help if more people understood this process. Unfortunately, it seems that a lot of people think science is like building an elaborate sculpture out of glass. If someone discovers that a piece of research is wrong, then it seems as if the whole sculpture cracks and falls to the ground. Creationists are particularly fond of this tactic. They seize on research about evolution that goes against earlier research, and claim that the entire theory of evolution is a fraud. They conveniently ignore all points on which scientists agree. So, for example, the researchers who have published the new findings on MYH16 do not conclude that humans were intelligently designed, MYH16 and all. Instead, they argue that the gene mutated earlier than once believed, and that the full history of this gene remains to be revealed. Science is more like a sculpture made of clay than glass, continually being molded and reshaped to better reflect reality.
Correction, 10/16/04: Changed "ancestors" to "relatives."
"But the upstream DNA--the part of the gene that could still be copied--told a different story. It showed signs of having undergone selection."
Can you elaborate on this (or point me somewhere, book or Internet, that explains)? What are the "signs of having undergone selection"?
Thanks.
Fantastic post, as is all of the writing on this site! I tried writing a non-technical description of GLUD2 which probably originated by retrotransposition during a time of hominid specialization. I ended up pretty tongue-tied- so now I REALLY admire your efforts!!
There are a number of ways of measuring natural selection on DNA. For example, one way takes advantage of the fact that only some of the "letters" in a gene are actually used to make a protein. So a mutation to this "coding DNA" can have a positive or negative effect on its owner, while a mutation to "non-coding" DNA is almost certain to have no effect. If a gene evolves without strong selection, you'd expect that the coding and non-coding DNA would have acquired equal levels of mutations. If selection is at work, the ratio will not be the same. So, for example, if natural selection favors new versions of the gene, you'd expect more change in coding DNA vs non-coding DNA. Scientists can measure these ratios by comparing versions of the same gene in different species.
The answers lead to more questions ...
Does a mutation to coding DNA ever have 'no effect'?
(Is it possible to tell?)
Does selection ever 'react neutrally' to a mutation? (for example, gene undergoes mutation, codes protein differently, organism as a whole unaffected in terms of interaction with environment)?
Thanks again. I have always loved biology, and I really appreciate this opportunity to learn :)
Yes, fantastic post & great blog, thanks so much. Maybe here instead of "the idea that chimpanzees are our closest living ancestors" you want to say "closest living relatives"? Regardless of how chimpy the common ancestors may have been, living chimps aren't ancestors.
Just a nit-pick here, but a truncated protein would be the result of a premature termination codon. That wouldn't send any DNA or RNA copying enzymes to a halt; instead, the nascent polypeptide chain being formed on the ribosome would stop synthesis prematurely.
Susan:
If the ratio of nonsynonymous to synonymous mutations (dN/dS) is significantly greater or less than 1, it's a sign of selection. A synonymous mutation is one that doesn't change the amino acid sequence of the resulting protein, so we know it's neutral. A nonsynonymous mutation changes the amino acid sequence, and may be beneficial, detrimental, or neutral. If nonsynonymous mutations are being preserved at a greater rate than neutral mutations (when dN/dS > 1), then natural selection must favor the nonsynonymous mutations, indicating positive selection. This means amino acid sequence is evolving in a particular direction. If nonsynonymous mutations are being preserved less often than neutral mutations (when dN/dS against, which indicates purifying selection. That means selection is keeping the amino acid sequence the same. When the rates are the same (dS/dN = 1), then there is no selection. The gene in question is therefore in a state of neutral drift, and probably doesn't do anything important, at least not anymore.
Oh, certainly. Most mutations will not have any effect. For one thing, there are synonymous mutations, which won't change the amino acid sequence. And then even if the amino acid sequence is changed, many (if not most) mutations won't affect the protein's function, since a lot of amino acids aren't critical, and others can be replaced with similar amino acids without any effect.
Oops, there was a little truncation there (premature stop codon?)
To recap:
dN/dS > 1: positive selection
dN/dS Oops, there was a little truncation there (premature stop codon?)
To recap:
dN/dS > 1: positive selection
dN/dS Oops, there was a little truncation there (premature stop codon?)
To recap:
dN/dS > 1: positive selection
dN/dS Oops, there was a little truncation there (premature stop codon?)
To recap:
dN/dS > 1: positive selection
dN/dS < 1: purifying selection
dN/dS = 1: neutral drift
Basically the same thing Carl already said, but I had to get nerdy about it.
Arrgh, it did it again. Someone doesn't like my "less than" sign.