Redrawing the Family Tree

David Pacchioli
March 01, 1995

"Trees," says Melissa Frye, "are our lives."

Frye, a graduate student in biology at Penn State, is not talking about maples, or aspens, or Ponderosa pines, although she'd certainly be happy to speak up for these towering titans of the kingdom Plantae.

She's talking about phylogenies—the spidery diagrams that trace the relationships between organisms, defining the evolutionary pathways that led from the undifferentiated protoplasm of four billion years ago to today's mind-boggling diversity of species. Phylogenies are the family trees of life. For Frye, and for other systematic biologists, determining the precise arrangement of their branches is the starting point for all biological inquiry.

"Knowing relationships," concurs Linda Maxson, head of the biology department, "is the first step. Once you know the relationships you can start asking interesting questions."

From Darwin until the 1970s, systematics, the study of biological diversity and its evolution, was a matter of morphology. Taxonomists, systematists specializing in classification, depended on similarities and differences in shape and form, as revealed both in living specimens and in fossils. But gross organic form can sometimes throw even the best taxonomist for a loop. It's a function of natural selection, the very mechanism of evolution.

Take the case of the Australian shrew. To the casual observer, this tiny, beady-eyed creature looks exactly like the North American shrew. In its genes, however, it is actually more closely related to the kangaroo than it is to its American "cousin."

The two shrews started out as very different animals, morphologically speaking. "Because they faced the same selection pressures," explains assistant professor of biology Blair Hedges, "they found the same niche and developed the same characteristics." It's called adaptive convergence, and it means the true evolutionary relationship of these organisms, their correct placement on the tree of life, has been obscured. "Natural selection," as Hedges somewhat ruefully concludes, "has frequently misled us."

Lately, however, systematics has found, in the advent of powerful molecular technologies, a check against this kind of subterfuge. The same DNA fingerprinting techniques that make the evening crime news are being used in systematics labs everywhere, to resolve the dilemmas that gross morphology cannot.

Don't tell Linda Maxson that the molecular approach is anything new.

"I've been doing it for 20 years," she says, flashing a subtle smile. Maxson has spent most of her career working with micro-complement fixation, or MC'F, a technique pioneered by her adviser Allan Wilson at Berkeley in the late '60s. MC'F makes use of a basic principle of the immune response. You take an antigen of a common protein like serum albumin, and raise antibodies to it. Do this for several species of organism, then test antigens against antibodies. Because an antibody reacts most strongly with its own antigen, you can use this means to tell species apart. By the relative strength of the reaction, you can tell how far apart genetically one organism is from another.

Wilson used MC'F to posit the recent divergence of humans and chimpanzees—which has been debated ever since. Maxson has used the technique, with considerably less controversy, to trace the phylogenies of amphibians. ("People don't get as passionate about the relationships between frogs," she says, then corrects herself: "Well, herpetologists do.")

Even earlier, another current Penn State biologist, Eberly professor Robert Selander, was among those who pioneered allozyme electrophoresis, a technique for separating the proteins in blood. "This," says Maxson, "was the first molecular phylogenetics."

These two techniques are now widely used, providing reliable, if relatively broad, estimates of genetic diversity. But molecular genetics really leaped forward in 1977, when it became possible to look directly at DNA itself.

Sequencing DNA, that is, isolating a gene that codes for a particular trait and identifying the precise ordering of nucleotide base pairs that make it unique, became a reality with the development of gene cutting techniques and cloning. (Cutting techniques allowed the isolation of small regions of DNA; cloning allowed these regions to be copied until there was enough material to work with.)

Early sequencing techniques, however, were difficult and time-consuming. Cloning, especially, was a convoluted laboratory procedure that took weeks.

Then, in the 1980s, a couple of key developments came along. First, Kary Mullis, a technician at a California biotechnology firm, discovered the polymerase chain reaction, PCR, a simple technique which amplifies DNA without cloning. Then came the discovery of Thermus aquaticus, or TAQ, a bacterium that dwells in hot springs, whose capacity to survive high temperatures makes its enzymes ideal for catalyzing PCR.

By 1990, PCR had been named "molecule of the year" by the journal Science. Now automated PCR machines are de riguer, and production of TAQ is a multi-billion dollar industry. Even small labs can get into sequencing.

The result is an explosion of work, as systematists rush to apply DNA technology to long-troubling evolutionary questions. A spate of new journals has appeared. There's been an influx of funding (albeit at the expense of more traditional taxonomic research), and there's a palpable sense of excitement.

In Blair Hedges' lab, Melissa Frye is tending to a gel-slab apparatus, two vertical panes of glass sandwiching a layer of clear conductive gel; a fresh batch of DNA samples, placed in the gel, is now being administered a low dose of electric current. In a couple of hours, this process will sort the molecular fragments by size.

The DNA in question this day is that of several species of the genus Amphisbaenia, a Caribbean worm lizard. Frye is only too happy to show off some whole specimens of this weird creature, pickled to ghastly hue and tagged in a jar she removes from a tall glass case in an adjoining room.

Across the lab two PCR machines sit side by side on a counter like twin retrievers, idle at the moment but ready for action. Hedges has given them names, affixed with tape: Corniche and Silver Shadow. "These are top-of-the-line machines," he explains; They cost $14,000 apiece, but to Hedges they are well worth the price. Fed a small sample of DNA, the machines will run through 30 or so cycles automatically, quickly producing enough genetic material for sampling. "They allow us to do in one hour what would take four hours in a standard machine."

Hedges has used this increased productivity well: Over the last few years he has published several papers in leading journals, on a number of different evolutionary subjects. Among the sticky questions he has addressed is the origin of birds.

"The fossil record suggests that birds evolved from reptiles, about 200 million years ago," Hedges explains. But morphological evidence and some early molecular results point to a closer relationship to mammals. "These conflicting results," he writes, "have created uncertainty about our ability to resolve amniote phylogeny." (Amniotes are land-living vertebrates which develop in an amniotic sac: mammals, birds, and reptiles.)

To resolve the dilemma, Hedges collected molecular-sequence data from 14 different genes across representative amniote species. When the sequences were compared, "we found that birds do indeed cluster strongly with reptiles.

"We made the paleontologists happy," he concludes. "There was no statistically significant evidence for this relationship before. Now there is."

Another puzzle that has attracted Hedges' attention is the closest lineage to the tetrapods—those animals with four limbs and their ancestors. Roughly speaking, this line runs from mammals back to birds, reptiles, amphibians, and at last to fish. "So another way of phrasing the question," Hedges says, "is 'Which fish is our closest relative?'"

Traditionally, that place has been given to the coelocanth, a stout, mottled creature that dwells off the east coast of Africa and was once thought to be extinct. The coelocanth, beneath its fins, has jointed, bony elements—the precursors of limbs. When Hedges, Maxson, and research associate Carla Ann Hass sequenced long stretches of the mitochondrial DNA of coelocanth and other fish, however, their subsequent analysis yielded a phylogenetic tree that favored another species: the lungfish, which can maneuver on land and breathe air. They published their preliminary findings in a letter to the journal Nature (June 10, 1993), but Hedges is not yet ready to consider this question answered. "We didn't have the highest degree of statistical confirmation," he notes. "We need more data."

Then there are the mammals. In the small conference room in Hedges' lab, over her lunch-on-the-run of Tastykakes and Coke, Frye offers a brief overview. "People think it's all figured out," she begins. "But there are big questions on which we have no clue.

"How are primates related to other orders? We don't know how the three big groups of mammals—the marsupials, monotremes, and placentals—fit together. We don't know when the monotremes diverged." (Monotremes are egglayers, like the duck-billed platypus.)

Mammals evolved, as Frye explains it, during an explosive era of change called an evolutionary radiation. In this brief (genetically speaking) window of time, between 80 and 100 million years ago, a few basic orders split into nearly 4,000 species, in order to take advantage of openings, as it were, in the Earth's developing ecology.

"It happened very quickly, and didn't leave a lot of fossils," she says. "So it's hard to reconstruct what actually happened."

How close are rabbits and rodents? How did bats evolve wings? "There's an argument that bat wings evolved twice," Frye reports. "Once for the small, common bats, and once for the larger fruit bats." Fruit bats, she explains, are thought to be closer to primates, because of similarities in the neuro-visual pathways of the two groups. "They use their eyes more than common bats do." Under a recent morphological argument, two rather distant species each developed wings separately, in response to similar environmental pressures. DNA sequence data suggests, however, that the two bats are a single group, that "Wings only evolved once," Frye says.

The bulk of Frye's own work thus far has been done on the guinea pig, Cavia porcellus.

In 1991, a researcher at Tel Aviv University, Dan Graur, suggested a split in the mammalian order Rodentia: specifically, that the group including C. porcellus represented an evolutionary lineage independent from your run-of-the-mill Rattus norvegicus or Mus musculus. Graur's result, based on an analysis of sequence data, stood in direct contrast to established morphology; and rodents, says Frye, "are one group for which we have good morphological data.

"It generated a controversy."

Graur had not, however, produced his own sequences in the lab; he had pulled inventoried samples from Genbank, a database administered by the National Institutes of Health. So Frye, under Hedges' direction, sequenced almost 3,000 base pairs of C. porcellus DNA, and compared it against a range of mammalian sequences. Her results strongly contradict Graur's, re-establishing, for now, the case for one big rodent family. Still, Frye reports, there were some stray guinea-pig genes among her samples that did not look rodent-like.

"Our theory is that this is due to the nature of guinea pig insulin," she explains. Guinea pig insulin can't bind zinc, which means it can't form the normal polymer for converting glucose. In short, it doesn't work. "It's terrible—only about 3 percent as effective as bovine insulin." So, then, how does the guinea pig survive? By compensating genetically.

"This might be the effect we're seeing. Maybe certain guinea pig genes had to evolve very fast to keep the animal alive."

In addition to answering basic questions of kin, molecular techniques can unravel complicated patterns of biogeography: the spread and distribution and division of species. Phil Heise, a graduate student of Maxson's, has used DNA sequencing to probe the evolutionary migration pattern of the Galapagos lava lizard, Tropidurus.

Since Darwin, the Galapagos islands, 600 miles off the coast of Ecuador, have been a classic place to study evolutionary migrations. The islands lie directly in the path of the Humboldt current, which pushes north from Antarctica until the contour of the South American coast redirects it to the west. For many species, including the famous tortoise, it is supposed that "colonization" happened this way: A pregnant female tortoise of a species native to Ecuador's west coast was swept out to sea, drifting far enough so that the current carried her clear to Galapagos. There she laid her eggs, and gradually, over succeeding generations, this original female's offspring made their way in the same fashion to the neighboring islands, where each line then evolved separately. A single lineage from the mainland spread and evolved into the numerous species that now exist.

"People assumed the same was true for Tropidurus."

His recent genetic work shows otherwise. Heise obtained samples of DNA of Tropidurus lizards from colleagues at the University of New Mexico and at the Los Angeles County Museum of National History. After sequencing a 300 base-pair segment of the sample, he was able to demonstrate, by comparing changes in the sequence of base pairs, that lizards of some island species differ more, genetically, from each other than they do from certain mainland species: They are more closely related to mainland lizards than to each other. Hence, Heise concludes, there were multiple colonizations of Galapagos by several mainland lizards.

By the time he finishes sequencing, Heise hopes to identify just which mainland species gave rise to the island populations, and also to reconstruct the historical order in which the individual islands were colonized. He also hopes to identify just how many species of Tropidurus there are on Galapagos. "There may be as many as a dozen."

Another of Maxson's graduate students, Jennie Hay, is working with another rare herp: the tuatara, ancient reptile of Hay's native New Zealand. Tuatara, Maori for "spiny-back," are the last of the Sphenodontids, an order that has survived largely unchanged since the Mesozoic period.

Tuatara have lasted this long because of New Zealand's relative isolation for 80 million years: in a land without indigenous mammals, these unblinking lizard-like creatures, which range up to two feet long and clench their prey with vise-like jaws, became one of the top predators. When humans and other mammals arrived a mere thousand years ago, the tuatara went into rapid decline. Today, the only ones remaining are scattered over 30 tiny islands off New Zealand's coasts, and their continued existence is exceedingly precarious. Because of their small size and the fragility of the ecosystem, Hay notes, "These populations could be very rapidly wiped out if rats or other predators are established on the islands."

An intensive study initiated in the late 1980s by Charles H. Daugherty and other scientists at Victoria University (including Hay, then a research assistant) undertook the first genetic survey of the tuatara, using protein electropohoresis. This work revealed the existence of two distinct species, falling into three distinct geographic groups—an crucial finding for conservation efforts. "Morphology wasn't adequate to tell them apart," Hay says.

Her tests at Penn State using DNA sequencing are aimed at revealing the tuatara's history. By comparing patterns of genetic variation in tuatara with that in some species of lizards on the same islands, she hopes to answer questions about the tuatara's evolution and spread.

A correct taxonomy, she notes, is essential to effective protection of this relic from the age of dinosaurs. "If you know what you've got," says Hay, "you know what needs most attention."

Up to now, most phylogenetic sequencing has depended on a few standard genes, mostly taken from mitochondrial DNA, a small loop of genetic material that lies outside the cell nucleus. The reasons are practical: the mitochondrial loop contains only 30 to 40 genes. The nucleus, by contrast, contains 100,000 genes, most of them quite complex, and lots of junk DNA besides. It's a lot harder to sort out.

Mitochondrial DNA is useful for many inquiries: because it evolves faster than nuclear DNA, it is especially good for tracking fairly recent evolutionary changes, like those that differentiate the larger mammals. With older divergences, however, like those separating Maxson's amphibians, you need large and slowly evolving segments of DNA to see significant change.

Maxson is currently working on establishing the nuclear gene for albumin, her long-time tool for MC'F, as a standard sequencing gene. "Albumin is present in all vertebrates," she says, "and it evolves at a rate that's good for looking at changes that have occurred in the last 100 to 150 million years." Bringing it into wide use "is a matter of developing the necessary primers," the short specific stretches of DNA that are applied to jump-start gene amplification.

There is also the matter of developing acceptance. Maxson remembers the reaction in the late '60s when her Berkeley mentor Wilson, based on MC'F work, posited a common ancestor for humans and chimpanzees between 3 and 5 million years ago. "Nobody believed him." Most investigators now accept Wilson's dates for humanity's divergence from the great apes, but there remains acute controversy among biologists over the use of genetic techniques, especially where morphological and molecular evidence conflict.

"Sometimes the debate is too caustic to reach print," says Hedges. He recently ruffled some feathers with a study reinterpreting the relations of web-footed birds.

The avian order Pelecaniformes, including not only pelicans, but boobies, frigatebirds, and cormorants, has long been distinguished by two unusual evolutionary features: a foot that is completely, as opposed to partially, webbed, and an expandable throat pouch, used for storing undigested food. The combination of these two features, Hedges and colleague Charles Sibley write, "has seemed so unlikely to evolve more than once that the monophyly of the group has been widely accepted" since the eighteenth century.

The problem is that the members of Pelecaniformes, in certain other characteristics—their pelvic musculature, the arrangement of their carotid arteries, and other, subtler, bodily features—vary considerably. Altogether, Sibley wrote in 1990, "this group may present the most complex and controversial questions in the avian phylogeny."

Hedges and Sibley decided to look at the molecular data. After sequencing mitochondrial DNA from 16 avian taxa, including the pelecaniform birds, they found that pelicans, boobies, and frigatebirds each belong in a different family. (Cormorants tend to cluster with boobies.) "There's no way they're all one group," Hedges concludes.

When he presented his data at a recent meeting of evolutionary biologists, however, "it was very tough for some of them to swallow. One leading ornithologist later told me, 'The samples must have been switched.' He could not accept these data."

Hedges and Sibley subsequently published their work in the Proceedings of the National Academy of Sciences, and they appended a strongly worded general suggestion for resolving conflicts between morphological and molecular estimates of phylogeny. Where consensus cannot be obtained, they state, the morphological evidence should be re-evaluated.

The reason for assuming the superiority of molecular evidence, as Hedges explains it, goes back to the idea of adaptive convergence.

With morphology, as in the case of the shrew, "It's very easy to have convergence," that situation where similar selection pressures lead to the evolution of similar—and misleading—characteristics. At the molecular level, however, because so much of the genome, the sum of an organism's genetic information, is meaningless redundancy or "junk," there's a great deal of variation that is completely unrelated to success or failure of the organism. Random mutations pop up frequently at nucleotide sites that don't code for anything. The likelihood of significant convergence at these sites—that they would independently evolve in exactly the same direction—is infinitessimal. In sum, Maxson says, "These molecules are exempt from the pressures of selection, so they can give us an independent—and extremely objective—estimate."

Molecular techniques, Hedges says, do not replace traditional approaches, but complement them. "With molecular data, you can take the real stuff of evolution—the color of a feather, the shape of a horn—put it on a tree based on molecular findings, and see the way things evolved."

Maxson puts it more strongly: "DNA is not just another measure to add to the others. It is the basic measure." Still, she adds, "People used to say a species is whatever a good taxonomist says it is," adds Maxson, "and 99 times out of a hundred the taxonomist was right. A good taxonomist has what [Nobel prize-winning biologist] Barbara McClintock calls 'a feeling for the organism.'

"It's that 1 percent that you get excited about. The molecular approach becomes exciting and invaluable where morphology isn't reflecting what really happened."

Linda E. Maxson, Ph.D., formerly professor and head of the department of biology, is now adjunct professor at Penn State and associate vice chancellor and dean at the University of Tennessee. S. Blair Hedges, Ph.D., is assistant professor of biology, 619 Mueller (865-9991). Melissa S. Frye, Philip J. Heise, and Jennifer M. Hay are doctoral students in biology.

Last Updated March 01, 1995