An ancient shrub that grows on only one island in the world has inexplicably swallowed up six mitochondrial genomes' worth of DNA, puzzling scientists and perhaps shedding light on the “junk” DNA that bloats the human genome.

For reasons scientists can only guess at, Amborella has ingested the entire mitochondrial genomes of a moss and three green algae plus several hundred miscellaneous genes from another alga and from an undetermined number of flowering plants. The hoard is orders of magnitude larger than what's been found in any other plant before. “The native genes are outnumbered by the foreign genes six to one,” says Jeffrey Palmer, a biologist at Indiana University Bloomington and corresponding author of the paper that reported the results in Science last December. “Nothing like this has ever been seen before. So that was utterly shocking.”

Dan Sloan, assistant professor of biology at Colorado State University who studies the evolution of mitochondrial DNA and was not involved in this study, says he found the results exciting. “It adds to a long list of amazing and downright strange things about plant mitochondrial genomes,” he says.

Mitochondria—organelles that convert food into energy cells can use—are the product of an engulfment of bacteria by ancient cells that would ultimately evolve into large multicellular life-forms. Mitochondria still retain evidence of their formerly independent lives, including their own DNA.

Amborella is an evergreen shrub with small white flowers that grows only on Grande Terre, an island in New Caledonia, a Pacific archipelago under France’s jurisdiction located more than 1,125 kilometers east of Australia. Amborella diverged from the rest of the flowering plants some 200 million years ago and is likely the first flowering plant that still survives today to have done so. That makes it an important organism for studying Darwin’s “abominable mystery” – the sudden evolution of flowering plants.

In this study scientists sequenced the mitochondrial genome of Amborella and looked for all the genes they could find. They discovered the near complete genome of a moss mitochondrion broken into four pieces scattered throughout the genome. The Amborella mitochondrial genome also contained the remains of whole genomes from three different species of green algae that are known to form lichens and small amounts of DNA from at least one more algal donor as well. There were also about two extra sets' worth of mitochondrial genes from what appear to be parasitic flowering plants. The vast collection has swelled Amborella’s mitochondrial genome to the size of a free-living bacterial genome.

Yet nearly all of this genetic haul seems to be nonfunctional. And it’s all been sitting there a long, long time. Judging by accumulated mutations, the transfers appear to have happened tens of millions of years ago. “It’s like Amborella is a museum full of ancient DNA that is just decaying away,” Palmer says.

Before the Amborella discovery the plant known to have squirreled away the most foreign mitochondrial genes was Rafflesia, a parasitic plant that makes the lurid-colored and notorious “corpse flower,” so called for its putrescent smell. (It also happens to be the world’s largest flower.) But Rafflesia contains 10 to 20 genes from a scattering of other flowering plants—nowhere near Amborella’s collection of 197 foreign genes and four near-complete foreign genomes, however.

How could this have happened? Amborella is often encrusted with mosses, lichens and parasitic plants, and the plant seems to be damaged often by falling branches or animals. Wounding breaks open cells and can allow the cell contents from two different species to mingle. In all plants, animals and fungi, mitochondria fuse with one another from time to time, which helps correct any mutations that may have arisen by allowing normal copies to be transferred to mutated mitochondria. Palmer and his colleagues propose that when Amborella and its epiphytes (plants that grow on other plants) were simultaneously wounded, mitochondria from mosses, lichen-derived algae and parasitic flowering plants mixed with the contents of Amborella cells that ultimately healed. Because all plants—and based on the evidence in Amborella, green algae—share the same method of mitochondrial fusion, the mitochondria of these unrelated species then fused with Amborella's and swapped DNA.

Amborella responds to wounding by making “suckers,” or new shoots, which may incorporate the cells bearing foreign DNA. Because most plant cells can theoretically become a stem-like cell, these cells have effectively been incorporated into a new germ line, or new individual plant. Such a process would be unlikely in animals because they sequester their germ cells—eggs and sperm—inside difficult-to-access ovaries or testes and never allow other tissues to reproduce.

But why is Amborella such a hoarder? No one is yet sure but it could be that the plant is not unusual in collecting so much foreign DNA. Instead, it might be comparatively defective in purging its mitochondrial genome. If so, Amborella seems to be perfectly fine even with its excess subcellular baggage. This could be interpreted as evidence that extra DNA is a neutral force in evolution, which would support the longstanding orthodoxy—questioned by recent research from the ENCODE (Encyclopedia of DNA Elements) Consortium—that the vast majority of the non-gene coding DNA in our own genome is also “junk” that serves no purpose but also carries little or no price.

Although Palmer felt Amborella’s unique flowering-plant heritage was merely a coincidence in this story, Sloan suggested its genetic stability over the 200 million years since its split from the rest of the flowering plants may have something to do with its reluctance to clean its mitochondrial closet. “Maybe Amborella is not so different,” Sloan says. “Maybe this actually happens a lot—the insertion of these large genomes into plant mitochondria. But because it’s more stable in its evolution, we can actually observe it better and it doesn't erase the footprints of those insertions as fast.”