Dec. 20, 2000 Penn State researchers have performed the most extensive study yet in an attempt to learn what the first flowering plants may have looked like. The genetic analysis was designed to find the first flower's closest living next of kin among 150 species whose genetic origins are thought to be the most ancient. The analysis revealed that the title of "oldest living flower" is shared by two very different-looking plants--water lilies and a rare woody shrub named Amborella.
A flurry of recent analyses of flowering-plant genes by other research teams had led scientists to believe that Amborella, which grows wild only on the remote island of New Caledonia in the southwest Pacific Ocean, was the winner of the title "oldest flowering plant," with water lilies coming in second. But because the Penn State team's results show both plants share first place, more options are now open for what the first flowering plant could have looked like.
The study, led by Claude W. dePamphilis, associate professor of biology at Penn State, was published in the 21 November 2000 issue of the Proceedings of the National Academy of Science. Knowing the precise genetic relationships among flowering plants would have "remarkable practical value and relevance," according to William L. Crepet of Cornell University, in a commentary on the Penn State research published in the same issue of the journal. Among the potential benefits are clues for the more efficient discovery of natural drugs, a precise framework for directing and evaluating the bioengineering of plants used for agriculture and medicine, and the ability to make more informed decisions about biodiversity conservation.
"Water lilies have one kind of flower that is bisexual, plus specialized cells called vessels for conducting water, but Amborella has two kinds of flowers--male and female--and the plant lacks vessels," dePamphilis says. If both plants are equally close genetically to the very first flowering plant, as his team's research indicates, then that ancestral plant has an equal chance of having possessed or lacked vessels and of having either bisexual or heterosexual flowers. Crepet agrees, stating "With both Nymphaea (water lilies) and Amborella at the base of the latest angiosperm tree, a wider range of characters might be expected in archetypal angiosperms (flowering plants) . . . than that implied by the previous consensus on Amborella alone."
First appearing on Earth during the age of the dinosaurs more than 140 million years ago, flowering plants, known as angiosperms, have been called one of evolution's greatest success stories and an important foundation of human society. Now the world's dominant form of plant life, flowering plants are the source of fruits, vegetables, grains, livestock feed, and medicines, in addition to comprising a large proportion of rain forests and other ecosystems.
According to dePamphilis, "Our estimate that both water lilies and Amborella are equally close to the root of the angiosperm family tree is an alternative hypothesis that needs to be entertained as seriously as the earlier studies." The researchers feel that the analysis methods they developed for the study not only give its conclusions substantial weight but also could serve as a model for further studies by other researchers.
They analyzed an extensive variety of leaves and flowers from plants they had been collecting from around the world for over a decade. "It has taken us ten years to collect tissues from over 1,000 plant species, including those whose genetic origins are thought to be the most ancient--plants growing as far away as Borneo and as close as right next to the lab," dePamphilis says. "We intentionally sampled these most basal lineages very densely to get the clearest possible picture of the tremendous diversity at the base of the angiosperm tree," dePamphilis says.
The dePamphilis team extracted and analyzed DNA from three separate compartments within the plant cells: the nucleus, where much of the cell's reproduction and protein-production occurs; plastids, which contain the cell's pigments; and mitochondria, which are involved in production of the cell's energy. "Molecular studies of plant evolution have mostly been done with plastid and nuclear genes, but recently our lab and a few others have started to use genes from mitochondria, which are known to evolve very slowly," dePamphilis says. "Several recent studies mixed mitochondrial gene sequences along with chloroplast and nuclear sequences, but we felt it would be important to analyze the three cellular components separately to guard against misleading results that could be caused by averaging them together," dePamphilis explains.
The researchers also are proponents of using more than one analysis method as another way of checking the accuracy of the analysis. "It wasn't possible to tell from previous studies whether the results were dependent on the method of analysis because, in most cases, only one method was used," dePamphilis says. So the researchers separately used three analysis methods--parsimony, neighbor joining, and maximum-likelihood phylogenetic inference--which construct trees of genetic relationships in slightly different ways. They used these three methods to draw family trees of their set of 150 plants based on relationships inferred by the similarity of their nuclear, plastid, and mitochondrial genes.
The team's approach was to analyze separately the DNA from each of the three cellular compartments with each of the three different types of analysis methods, predicting that concurring results would be a strong indication of a correct picture of evolutionary history. But they were surprised to find that they obtained different family trees for Amborella, depending on which type of phylogenetic analysis they used.
"When we did independent analysis of the mitochondrial DNA for Amborella, it didn't give the same result as did its nuclear and the choloroplast DNA," dePamphilis says. "We spent a good six months doing detective work to track down the source of that conflict and then trying to resolve it."
During that process, the team's postdoctoral fellow, Todd J. Barkman, discovered a previously unknown second copy of one of Amborella's genes, which turned out to be causing a lot of the conflicting results. "One mitochondrial gene, atpA, was placing Amborella much higher up the evolutionary tree than we expected based on other studies," says Barkman, who then discovered the atpA gene contained a pattern of "noisy" segments. "I found a minor signal in the sequence data for Amborella, which ran throughout the length of the atpA gene sequence. This signal turned out to be our clue that Amborella has two copies of this gene, and further experiments separated the two copies clearly." The team's reanalysis, taking the duplication into account, placed Amborella in its expected position at the base of the family tree of living flowers.
This discovery also led the way to an important addition to the overall analysis method the researchers eventually developed. "We are recommending that other researchers find and remove all excessively noisy components of their data, like the duplicate mitochondrial gene in Amborella, before they even begin their analysis," dePamphilis says.
To do this genetic housekeeping task, the Penn State team used a software package called Relative Apparent Synapomorphy Analysis (RASA), which was created by James Lyons-Weiler, a member of the team, who makes it available to other researchers at no charge on the World Wide Web. "RASA allows us to examine the data before we do any of the phylogenetic analyses so we can identify and remove any individual sequences that might mislead the results," dePamphilis says. Such "noisy" sequences typically contain excessive errors compared with the rest of the data. After cleaning up their data first with RASA, the team's analyses were in agreement about the root of the angiosperm tree for all three analysis methods and for all three cellular compartments.
The Penn State researchers feel they have developed a more powerful analytical protocol for getting more accurate estimates of any given piece of evolutionary history. "The strongest conclusions can be drawn from congruent results based on clean data from different sources of genetic material analyzed with different analysis methods," dePamphilis says. "We believe the conclusion of our study is strong, and indicates that there was a great deal of initial diversity among the earliest ancestors of today's flowering plants."
In addition to dePamphilis at Penn State, the research team also included: Barkman, who recently became an assistant professor at Western Michigan University; Lyons-Weiler, who now is now an assistant professor at the University of Massachusetts at Lowell; Gordon Chenery, a high-school teacher in Nashville, Tennessee, who earned a master's degree in biology under the direction of dePamphilis; and Joel R. McNeal, a graduate student at Penn State.
This research received financial support from Penn State and the National Science Foundation.
Other social bookmarking and sharing tools:
Note: Materials may be edited for content and length. For further information, please contact the source cited above.
Note: If no author is given, the source is cited instead.