NATURAL DIRECTION

Extraordinary new findings suggest that life forms may literally
direct their own evolution

When molecular biologist John Cairns arrived at Princeton for a debate in the
summer of 1990, the mood was tense. The arena, a lecture hall at the
university's esteemed Lewis Thomas Laboratories, boasted a wall-length
chalkboard, overhead slide projectors, and comfortable seats with armrests for
taking notes. But despite these academic trappings, Cairns could almost hear
"the saloon doors swinging, the train approaching, and the wind rustling down
the plains." Some of the spectators awaiting the debate had even dubbed it "the
shoot-out on Main Street. "

On one side of the dusty scientific road stood Bruce Levin, a professor from the
University of Massachusetts at Amherst. Levin, like almost all biologists,
believed that one elegant mechanism could explain the diversity of life on
Earth. According to this prevailing view, all species evolve through random
mutation of the genes. Populations with new traits arise when mutations produce
organisms especially good at finding food, avoiding predators, and producing
offspring. After generations, these successful mutants may replace earlier
organisms within the species or even form whole new species. The process is
called natural selection, since nature itself apparently selects the individuals
most likely to survive.

Convinced of this scenario, Levin had come to Princeton, scientific sixguns at
the ready, to stand down the heretics-Cairns and his colleague, University of
Rochester molecular evolutionist Barry Hall. Not only had these two renegades
challenged the prevailing orthodoxy, they had done so in Nature and Genetics, a
couple of the most prestigious scientific journals in the world.

Cairns and Hall were not creationists who believed that people had been placed
on Earth fully formed by heavenly design. Instead, they had come to Princeton
with an alternative scenario for how evolution works: The mutations that drive
evolution, the researchers claimed, were not always random. In experiment after
experiment, they said, microorganisms seemed to be whipping up their own
mutations-almost as if some inner molecular composer were helping the cells
react to environmental requirements and needs. They even had a name for this
shocking and powerful phenomenon: directed mutation.

At the Princeton conference, Levin argued against this radical view. "Mutants
arise at random," he said, "whether they are favored by natural selection or
not." Only after the mutants arise randomly, he added, does the environment kick
in, with natural selection acting as the "editor of evolution," choosing the
life forms it likes best among those already around.

Levin pointed out the technical limitations of the research, insisting that
Cairns and collaborators had merely assumed directed mutation was occurring
without sufficiently ruling out other, less radical explanations. To emphasize
what he saw as the work's major flaw, he titled his talk "Refrigerator Lights
and the Limits of Inductive Inference." The idea, he told the audience of
professors and graduate students, was that no matter how many mundane
explanations the researchers eventually disproved, there would always be more.
Their approach, he stated, was a bit like trying to prove that a refrigerator
light is off when the door is closed. "Even if you put a little kid in the
refrigerator to tell you what's going on," he said, "you could never be sure
the kid was telling the truth." They would never validate their theory, he
concluded, unless they found the mechanism at its root.

But despite these objections, Cairns and Hall were impossible to ignore. As one
seminar observer, Princeton graduate student Karen Weiler, recalls, "A lot of
people there that night wanted to dismiss Cairns and Hall, but they just
couldn't."

With good reason. The new research, if correct, would alter one of the most
entrenched scientific theories of our time, in the process changing our notion
of how life on Earth evolved. It might also explain the huge gaps in the fossil
record-long epochs during which paleontologists can find no evidence of
evolution at all. After all, if mutations are literally "directed" by life
forms reacting to environmental change, then rapid evolution would occur
primarily at highly "punctuated" moments-during ice ages, say, or meteor hits,
when environmental stress is especially great. In fact, if the new scenario
turns out to be valid not just for microorganisms, but also for more complex
living things such as rain forests, animals, and humans, evolutionary biologists
would have to rewrite much of the work they have labored over for the past
hundred years.

Today's evolutionary biologists-often called "Neo-Darwinists"--base much of
their work on the ideas of the master, Charles Darwin himself. While exploring
the flora and fauna of South America and the Galapagos Islands in the 1830's,
Darwin observed the immense variety of life. Even in a given species, there was
large variation from one individual to the next. Based on this, Darwin proposed
a brilliant theory for how evolution works: Nature was always generating
variation, he declared, and in the brutal struggle to survive, some variants
would just be more successful than others. Those better at exploiting the
environment, he said, would have more offspring, and these individuals would
prevail.

But despite this central insight, Darwin still didn't know why the variation
occurred. The reason: The world had not yet heard of the tiny hereditary units
called genes.

In the century after Darwin proposed his theory, however, biologists discovered
that genes, found in every cell, determined the nature of living things. By
orchestrating the synthesis of organic chemicals into the stuff of life, genes
dictated virtually every biological characteristic from brain size to eye color
and body type.

The genes themselves were composed of a helical molecule called deoxyribonucleic
acid, or DNA. All DNA, in turn, consisted of just four chemical building
blocks, called bases. These bases, strung together over and over like beads,
could be arranged in literally millions of combinations, creating the potential
for virtually infinite genetic diversity on Earth.

For a species to generate this diversity, said the Neo-Darwinists, all that was
needed were some simple chance events. In the random shuffle of life, a few
bases would accidentally be replaced by others. Over time, the accumulation of
such changes-called " point" mutations because they occurred one base at a
time-would result in different types of creatures, even whole new species.

The Neo-Darwinists who proposed this grand synthesis of Darwinian evolution and
modern genetics said their theories were rooted in scientific fact. The first
study to back them up was published in the 1940's, after Salvador E. Luria, then
of Indiana University, began wondering how he could prove that mutations had
occurred randomly and not in response to environmental stress. It occurred to
Luria that he could compare genetic mutation to another rare event-hitting the
jackpot at a slot machine in Las Vegas. People playing these slot machines
usually came up empty-handed, Luria knew, but every so often, by chance, someone
struck it rich.

Luria compared the slot machine to a colony of bacterial cells. Each cell
reproduced by dividing in half. The two resulting daughter bacteria, in turn,
reproduced by dividing in half again, and so on and so forth until, within a
couple of days, one cell had become a swarming bacterial colony of 1 billion
cells or more. If a cell were to mutate randomly, early in the life cycle of a
colony, it would produce large numbers of mutant descendants, resulting in a
"jackpot" of mutants.

Using this concept as the basis for experiments, Luria and his colleague,
Vanderbilt University physicist Max Delbruck, grew bacterial populations in test
tubes. Then, after the populations had grown, the scientists introduced a
lethal virus.

Mutants never seemed to emerge directly in response to the virus. Every so
often, on the other hand, a given population just happened to contain huge
numbers of mutants resistant to the virus. These mutants were so numerous they
had obviously arisen early in the life of the population, way before the virus
had been introduced, and represented jackpots of enormous proportion. The
conclusion: Mutants resistant to the virus occurred randomly, without input from
the environment. The environmental stress-in this case, exposure to a
virus-came into play only later, selecting out the mutants that could survive.

If the Neo-Darwinists were happy to see their theories boosted in this way, they
were even more overjoyed when, in the 1950's, geneticist Joshua Lederberg drove
the point home. Lederberg started with a gel containing numerous colonies of
bacteria. Then he pressed the gel onto a strip of velveteen as if he were
printing on paper with a rubber stamp. Finally he took the velveteen and
pressed it onto a second gel. When he pulled the velveteen off, the pattern of
bacterial cells on the second gel mirrored that of the first. To do his
experiment, Lederberg exposed only the bacteria on the second gel to the virus.
A certain number of cells were immune to the virus, and only they survived.

The question was, did the resistant cells-mutants-develop in response to the
virus, or were they there beforehand? To find out, Lederberg tested plate number
one, and voila: He found mutant cells resistant to the virus at the same exact
site as on plate two. Obviously, the mutant cells had been there all along.
Like Luria and Delbruck, Lederberg had validated the ideas of the
Neo-Darwinists: Mutant organisms, he showed, emerged spontaneously, without any
stimulation from the environment at all.

Satisfied with the evidence, the Neo-Darwinists spent the next 30 years refining
their theories, coming up with all sorts of situations under which populations
might evolve. But while they spent enormous time honing and polishing their
theory, the overriding mechanisms-random mutation and natural selection-remained
the same.

And there the matter stood until the 1980's, when it fell under the scrutiny of
British oncologist and molecular biologist John Cairns. A deeply thoughtful man
with the regal good looks and bearing of actor Peter O'Toole, Cairns had spent
years as director of the prestigious laboratory at Cold Spring Harbor in New
York and was now in Boston at Harvard's School of Public Health. Interested in
the mutations that induce cancer, he thought he might gain some helpful insight
by studying mutations in bacterial cells.

Naturally, he began by going back to the old studies conducted by Luria,
Delbruck, and Lederberg decades before. Examining their work more carefully, he
realized that though they had conclusively proved the existence of random
mutation, they hadn't ruled out other evolutionary mechanisms as well. Indeed,
their crucial studies were plagued by an overwhelming flaw: the use of a lethal
virus.

The virus presented a problem because, in bacterial populations, mutants
resistant to virus take several generations to express themselves. The reason
is that bacteria replicate by dividing and then growing. A first-generation
mutant thus contains half of the cellular material from the nonmutant parent
cell; in fact, new mutants carry so much parental material that they often seem
to behave like the original strain. Only after many generations, when the
original nonmutant gene products have been diluted out, can the mutant's new
characteristics truly emerge.

Therefore, in the Luria and Delbruck experiment, the very virus that might have
caused the production of resistant mutants would have killed those mutants
instantly. Mutants produced in direct response to a lethal virus would never be
observed.

To get around this problem, Cairns decided to see if he could generate mutants
through a less instantly lethal form of selection: He would deny his bacteria
access to all nutrients except for one that they lacked the ability to use.
Either they would learn to use the new nutrient, or else they would eventually
starve to death. "The question was," Cairns explains, "could some mutants arise
as a result of pressure from the environment?"

He began his experiment with populations of bacteria unable to digest the sugar
lactose. Then he placed these bacteria in a medium that contained only lactose
for a food source. Of course, the bacteria stopped multiplying because they had
no usable food. But after a few days, large numbers of the lactose-utilizing
mutants began to appear. The mutants were so numerous, in fact, that they could
not be accounted for by the theory of strictly random mutation. The suggestion:
The bacteria were learning to generate their own, useful mutations through a
surprising evolutionary process that wasn't random at all.

Cairns published his study in Nature in 1988. Near the end of the article, he
suggested some ways in which the environment could influence genetic material,
thus allowing directed mutation to occur. Each of these suggested processes,
Cairns had the chutzpah to write, "could, in effect, provide a mechanism for the
inheritance of acquired characteristics."

The statement inspired sentiments of fear and loathing among evolutionary
biologists worldwide. The term acquired characteristics, after all, smacked of
the discredited eighteenth-century biologist Jean-Baptiste Lamarck, who proposed
that evolution proceeded as individuals used various organs, muscles, and limbs.
For instance, Lamarck had declared, if a creature under stress was forced to
exert extreme muscular strength, offspring would inherit-or acquire-larger
muscles whether or not they actually required the additional strength. Cairns
had used the phrase " acquired characteristics" by way of analogy only; he was
talking about genes and proteins, not fingers and toes. But that didn't stop
his critics from writing to Nature in droves. They insisted that Cairns tighten
his laboratory controls and proposed alternative scenarios that would leave the
Neo-Darwinian interpretation intact.

But Cairns stood philosophically firm: "It's easy to imagine molecular
mechanisms that might drive the process of directed mutation," he explained.
"We've already proven feedback between organisms and the environment; this
occurs through messenger molecules that help genes communicate with the cell and
the outside world." In light of this, he added, "It seems almost perverse to
maintain, as a matter of principle, that evolution is driven only by random
mutations, and that no other phenomenon comes into play."

One researcher wholeheartedly agreed. Molecular evolutionist Barry Hall had
been on a similar track for years. His involvement in the field began in 1970,
while visiting his good friend, University of Minnesota population geneticist
Dan Hartl. Hartl had been studying the fruit fly Drosophila, monitoring how
large groups of these creatures evolved from one generation to the next. Hall,
on the other hand, was studying the general molecular and cellular biology of
the popular laboratory bacteria E. coli. "We got to talking," Hall explains,
"and kind of said, Gee, wouldn't it be nice if you could watch evolution as it
happened, on the molecular and cellular level, by experimenting with bacteria?"

A couple of years later, Hall began the work. He started with a strain of
bacteria normal in all respects but one: The individuals in his colonies lacked
an enzyme necessary for digesting lactose. He plated bacteria from this strain
on a dish containing a blood-red gel known as a Macconkey medium. Dissolved in
the medium were two types of food sources: a small amount of peptide and a large
amount of lactose. The gel was an important indicator, since bacteria that
digested lactose would absorb some of the dye, showing up as red; those that
digested the peptide would not absorb the dye and would thus appear white.

The bacteria, unable to digest the lactose, consumed all the peptides. As they
grew, they peppered the blood-red expanse of Macconkey with white colonies.
When all the peptides were gone, bacterial growth seemed to stop. But out of
curiosity, Hall let these seemingly stymied colonies sit around his lab for a
week or two. In every case, he found, pimples of red began bursting through the
islands of white. These red bursts, called papillae, were new colonies of
bacteria, now able to utilize the lactose. In short, they were mutants.

For nearly a decade, through stints at the University of Newfoundland and the
University of Connecticut, Hall watched his bacteria give rise to mutant
offspring capable of digesting lactose. As he performed the experiments, he
began to realize the oddness of his results. Time and again, his bacteria were
evolving the ability to eat the lactose about a hundred million times more
frequently than would be expected if mutation had occurred purely by chance.
What made the results especially strange was the magnitude of the genetic change
involved. Sequencing the bacterial genes, Hall discovered that two genetic
mutations, not just one, were required for digestion of lactose to occur.

Hall discovered the phenomenon in other E coli populations as well. He was
absolutely floored, for instance when he used his technique to create E coli
mutants that could thrive on the carbon source citrate. "This was weird," he
says, "because one of the definitions of E coli, one of the things that's used
to distinguish it from all other closely related organisms, is that it cannot
use citrate." Mapping the genes of his citrate mutants, Hall found "the
improbable stacked on top of the highly unlikely" when it turned out that the
citrate-consuming E coli had two large-scale genetic mutations, not just a
single altered base. The finding was so completely out of line with results
predicted by accepted evolutionary theory that Hall didn't know what to think.
"At that point," he recalls, "all I could do was throw up my hands." Yet by
1988, when Cairns described the phenomenon of directed mutation in Nature, Hall
realized that he had been studying this phenomenon as well. By then at the
University of Rochester, he had witnessed directed mutation in thousands of
bacterial colonies and had charted its course in many specific E coli genes. He
was also beginning to study the phenomenon in yeast.

Discussing the research today from his immaculate Rochester office, his spanking
new lab overflowing with projects next door, Hall expresses awe at the mysteries
he has seen. "For almost fifteen years," he says, "I have been slapped in the
face with the highly improbable. When that happens, you either get religion and
say, 'God is favoring me,' or you conclude that perhaps your understanding of
the process-in this case, the process of evolution-is incomplete."

Hall did the latter. Paying attention to his organisms, the lowly bacteria, he
has been able to reach just one conclusion: "While some mutations may be random,
many others are generated by the organism to cope with environmental stress."
Because these mutations are literally selected by the organism while it is under
stress, Hall calls them "selection induced."

To date, Hall has generated selection-induced mutations for half a dozen E coli
genes and a couple of yeast genes as well. Most of the time, he worked with
bacteria unable to utilize nutrients such as lactose. He has also worked with
bacteria unable to replicate because they lack the ability to manufacture
critical amino acids, the building blocks of protein. When he first places
these bacterial strains on a plate or in a liquid medium, the cells seem to stop
growing. But after a few weeks, Hall finds large numbers of mutants that can
utilize the nutrients or manufacture the needed amino acids.

In dozens of control studies, moreover, Hall has shown that the mutants are
specific to the environment. The starving cells do not just start churning out
mutants at random. If lactose is the only nutrient available, for instance, the
mutants will develop the ability to digest only lactose, not some other,
unrelated sugar. If the medium is missing the amino acid tryptophane, then the
cells will evolve the ability to produce that amino acid only.

These days Hall and Cairns regularly correspond. One of their most pressing
concerns: Figuring out how bacteria and yeast can possibly "know" what mutations
to make. As Hall himself says, "It's implausible that a single cell has an
array of machinery complex enough to measure the environment and then, in
effect, say, 'Oh, this is how I have to mutate,' and then just go out and do it.
Yet that is what seems to occur."

No viable theory has yet emerged, though Cairns and other researchers have
speculated on the existence of something like a spontaneous mutation generator.
"Imagine," says Cairns, "that these guys [the bacteria] are out there
struggling, and they're not multiplying because of the stress. Mutations are
spun out and then gotten rid of, until finally one is good. The lights go on,
the dynamo starts humming, and the cell can grow. At that point the mutation
generator comes to a halt."

No matter what the mechanism, however, one question dominates: Even if this
eerie phenomenon plays a large role in the evolution of microorganisms, does it
have a similar impact on the human species? Evolution, Cairns believes, works
the same way for the simplest, one-celled organisms and the most complex. No
matter what the life form, he says, "the process is the same." Adds Hall, "As
organisms evolve, they affect the environment. The environment, in turn, has an
impact on life. If directed mutation turns out to be a powerful evolutionary
force, we may have to reanalyze the feedback loops between the biosphere and the
earth."

But Levin of Amherst insists that, despite the elegance of some of the work, it
is not strong enough to stake a claim. "Until Cairns, Hall, and others show the
mechanism by which directed mutation takes place, I will be skeptical," he
states. "They certainly haven't shown that organisms perceive the environment
and then understand what they need, nor have they demonstrated that organisms
have the cellular machinery for this perception."

Some of the strongest criticism to date has been offered by evolutionary
biologist Richard Lenski of the University of California at Irvine. Working
with graduate student John Mittler, Lenski has recently published a paper in
Nature himself. According to Lenski, cells may simply generate large numbers of
certain types of mutations when they are starved, as Cairns and Hall's cells
are. He also suggests that some bacterial populations may increase in number by
literally consuming bacterial waste products; with more cells in the population,
one might expect to find a larger number of mutants.

Hall, for his part, counters that he continues to test all possible explanations
for directed mutation in his lab; as critics suggest additional control
possibilities, he says, he will test those as well, "no matter how foolish they
seem." None of the explanations posed so far, he adds, come close to explaining
the effect, at least according to his painstaking control studies in the lab.
To bolster his argument, he takes out a stack of papers currently in press and
reams of data from his shelves. Drawing furiously on his chalkboard, he seems
to demolish the notion that cell starvation or an undetected increase in colony
size can account for the numbers and types of mutations he has seen.

Cairns, soon to retire to his native England, says that the critics "see
themselves as crusaders defending some religion, and by hook or by crook, defend
it they will. But the world will pass them by." The reason, Cairns notes, is
the power of science itself. "Our studies are ever more detailed," he says,
"and system after system seems to be demonstrating this effect. The data will
speak for itself."

If that data holds, evolutionary biologists will have to go back to the drawing
board and rewrite their theories of how earthly life evolved. If directed
evolution turns out to affect not just microorganisms, but also more complex
living things, then we may have to reanalyze the fossil record and revamp the
history of Homo sapiens as well. Says Hall, "It would require a paradigm shift
in the way we view the world."

Whichever way the evidence finally points, however, it will be business as usual
for lab hound Barry Hall. "The Neo-Darwinists claim that evolution works too
slowly, and on such large populations, it's simply impossible to study the
process," he concludes. "But for people working with bacteria it is possible to
study evolution as it happens. Biology is an experimental science, not a
theoretical one. The business we're involved in is asking-not telling-the
universe how it works."