Stable Across Time and Species, Primeval DNA May Be Chromosome Counter

Ultraconserved Elements Found Absent from Commonly Duplicated or Deleted Segments

The most seasoned geneticist may feel a tinge of awe when encountering the basic fact that less than three percent of the human genome codes for actual proteins. Researchers have been avidly scanning the remaining territory, the so-called dark matter of the genome, for signs of life-promoting function. Three years ago, a team of scientists made a stunning find. Buried in the chromosomes of every vertebrate were hundreds of stretches of DNA so stubbornly unchanged that they could be the genetic equivalent of living fossils. The coelacanth, a primitive fish found in 1938 in waters off the east coast of South Africa and thought to exist relatively unchanged since its origin more than 400 million years ago, carries some of these ultraconserved genetic sequences, and in a form largely similar to those in humans.

Photo by Graham Ramsay

“With segmental duplications and copy number variants, the observed overlap with ultraconserved elements was far below anything you would expect,” said Adnan Derti (right). He is shown with Chao-ting Wu (left) and George Church.

Though the majority of these stretches are located in the intervals between protein-coding regions, some of the ultraconserved elements (UCEs) overlap such coding regions, in particular those involved in development. This finding has led to speculation that UCEs may act as regulatory agents, such as enhancers. Yet even if they were serving a dual function, regulating and coding, it is unlikely that they would have resisted even a single base change. “It is very unlikely that a protein or RNA structure would require that level of conservation,” said Chao-ting Wu, HMS professor of pediatrics (genetics). How have these identical sequences, at least 200 base pairs or more in length, resisted the force of random mutation and natural selection for so long?

Wu, Adnan Derti, and colleagues have hit upon a tantalizing clue. Running millions of computer-generated matches, they found that these primeval sequences are strikingly depleted in stretches of the genome that have undergone segmental duplications or deletions. On this basis, the researchers propose that UCEs may be involved in carrying out one of the most essential, and earliest, of cellular activities—ensuring that a diploid cell has no more than two copies of each chromosome.

“One possibility is that UCEs might be active during fertilization, that there is some kind of check at that point, to make sure two compatible genomes have been brought together and that it is worthwhile to invest in this offspring,” said Derti, HMS research fellow in biological chemistry and molecular pharmacology. The findings appear in the October Nature Genetics.

Swappable Blocks
What has impressed some is the way the study pairs UCEs with another hot discovery in genetics. Humans were long thought to vary from one another by small base changes, but it now appears that they carry much bigger variations. Whole chunks of DNA may exist in duplicate, triplicate, or higher multiples in some people—and not at all in others. “The differences are so great between you and me that it would be no surprise if we differed in 20 places,” said Wu. Some of these copy number variants (CNVs) may give their bearer a slight adaptive edge and, by natural selection, spread through the entire species, at which point they are labeled segmental duplications. Derti, Wu, and colleagues found that UCEs were absent in CNVs, as well as in segmental duplications.

“If you had to pick something from the human genome project that was new and exciting, other than new tools, it would be these two things, UCEs and CNVs,” said George Church, HMS professor of genetics, who, with Frederick Roth, HMS assistant professor of biological chemistry and molecular pharmacology, contributed to the study.

Two to Tango
This timely convergence began when Wu came across the 2003 paper by University of California, Santa Cruz, researcher Gill Bejerano and colleagues announcing the discovery of UCEs. Among their myriad observations, the authors mentioned that UCEs were absent from both the Y chromosome, which typically exists in only one copy, and chromosome 21, which can exist in triplicate in humans. Wu, who has a long-term interest in the pairing of chromosomes, was immediately struck. “I began to wonder, maybe UCEs were copy counting in the strictest sense of the word. They are there to make sure they are present twice and only twice in each diploid cell,” she said. She and Church began looking for a way to test the copy-counting hypothesis.

Meanwhile, Derti had been working on a gene involved in alternative splicing that turned out to be overlapped by a UCE. Intrigued by the mysterious sequences, he decided to join Wu and Church. They began by taking a set of nearly 900 UCEs and, using computer-based techniques, matching them to a known group of CNVs. If UCEs were truly counting themselves for the good of the cell, one would not expect to see them in duplication-prone regions. That’s not what Derti found—there appeared to be some overlap between the UCE and CNV sequences.

He decided to switch his attention to segmental duplications, which are evolutionarily older and might have had a chance to rid themselves of potentially troublesome UCEs. Sure enough, the UCEs did not overlap the group. It turns out, the original set of CNVs actually combined data from two different sources. He went back, this time separating out the CNVs according to source. One set of CNVs exhibited a strong avoidance of UCEs. As more CNV datasets came out, he tested those and found the same avoidance. Just as segments of DNA can be duplicated, they can be deleted in some people. Derti tested a dataset of 1,000 segmental deletions from the labs of David Altshuler, HMS associate professor of genetics, and others. They, too, exhibited a lack of UCEs.

“So the model right now is, a UCE is unlikely to be deleted or duplicated—or if it is, it is unlikely to survive,” said Wu. “What is compelling for us now is, how did UCEs get ultraconserved?” One possibility is that UCEs somehow resist rearrangements. “Something about them prevents that,” she said. “Or when they happen, they are immediately repaired so natural selection does not have a chance to work on them.” Another is that duplications and deletions occur, but are eventually selected against. If that were so, then they should be found in some CNVs, which can be evolutionarily quite recent.

And this still leaves open the question of why the UCEs have traveled down through the ages completely unchanged. The copy-counting hypothesis addresses this issue, Wu said. Though it is not clear how UCEs might carry out such a function, it is possible that maternal and paternal UCEs recognize each other on homologous chromosomes, perhaps by a physical pairing mechanism. In some cases, deviations by even one base may disrupt the fit, which could be deleterious for the cell.

The researchers emphasize that the copy-counting hypothesis is just that—a hypothesis. Nor do they contradict the idea that some UCEs act as enhancers of nearby developmental genes. Bejerano believes that their discovery that UCEs are depleted in duplicated regions of the genome strengthens this view. “It suggests duplication of these regions may interfere with the transcriptional program, perhaps very early during development,” he said.

Taken as a group, UCEs may carry out multiple tasks, including a more global stabilizing function. “When you bring your car for an inspection, they check that you have four tires and doors, and so forth, but the shapes of cars may be very different,” said Derti.  “I think it is the same in the genome. There is an advantage to being able to adapt and to have new traits and to have genes being copied. But the fundamental recipe for making an organism, say a vertebrate, has to be stable. And that is where, I think, these UCEs come in. So part of the genome has to stay fixed, but around that, the genome can be modified.”

—Misia Landau

top