Chapter 88 15.6 Dumb Scientists Evolving Smart Molecules

Artificial evolution isn't limited to silicon wafers, though.Evolution can be introduced wherever engineering methods fail.The cutting-edge field of bioengineering has embraced synthetic evolution. This is a real world question.You need a drug to fight a disease whose mechanism has just been isolated.Think of this disease mechanism as a lock.All you need is the right key - a drug - to open the lock. The composition of organic molecules is very complex.They consist of thousands of atoms arranged in billions of ways.Just knowing the chemical composition of a protein doesn't do much to help us understand its structure.The long chains of amino acids are layered on top of each other, and the hot spots—the active parts of the protein—are just in place on the outside.The way the protein folds is like twisting a mile-long rope with six dots marked in blue so that the dots all fall on different outer faces.There are countless ways of winding, but few meet the requirements.You don't even know if a way is close to the answer—unless you're almost done with it.There are so many variations that it would be impossible to try them all, even if they were exhausted.

Drugmakers generally have two tools to deal with this complexity.In the past, pharmacists relied on chance.They tried all the chemicals found in nature to see which one would unlock a given lock.There are usually one or two natural compounds that work in part - and that's part of getting the key.Today, in the age of engineering, biochemists try to decipher the pathways between genetic code and protein folding to see if they can engineer the steps needed to build molecules.Despite some modest successes, protein folding and gene pathways remain too complex to control.Thus, this logical approach, known as "rational drug design," has actually hit the limits of the complexity that engineering methods can handle.

Since the late 1980s, bioengineering labs around the world have been working to perfect another tool we use to create complex bodies—evolution. Simply put, the evolutionary system generates billions of random molecules that are used to try and pick locks.Among these billions of trivial candidates, perhaps only one part of a molecule coincides with one of the six points of the lock.This "affinity" key was kept, and the rest were ruthlessly eliminated.Billions of new variants are then bred from the surviving "affinity" keys, while the point where the lock fits remains unchanged (called a binding), and are used to try that lock.Perhaps at this time another "affinity" key can be found that can match two points.This key is kept as a survivor, and the rest die.Survivors bred billions of variants, and the best-fit offspring survived.Over several generations of this knock-out-mutate-bind process, the molecular breeding program finds a drug—perhaps a lifesaver—that fits all the points of the lock.

Almost any kind of molecule can be evolved.For example, biotechnologists could evolve an improved version of insulin.They injected insulin into the rabbits, and the rabbit's immune system developed antibodies (antibodies are the toxin's complementary form) to the "toxin."Next, extract this antibody and inject it into the evolution system.In an evolutionary system, antibodies are like a test lock.After several generations of evolution, biotechnologists could arrive at a complementary configuration of the antibody, which is, in effect, a substitute version of insulin.This alternative is extremely valuable.Alternatives to natural medicines offer many advantages: They may be smaller; more easily injected into the body; have fewer side effects; be easier to manufacture; or be more precisely targeted.

Biological evolutionists can also evolve an antibody against hepatitis virus, and then evolve a mimic hepatitis virus that matches the antibody.The selected virus is not a perfect variant, but lacks some active points that can cause fatal symptoms.This defective, incompetent substitute is what is known as a vaccine.Thus, vaccines can also be evolved through non-engineering methods. All the usual logic of drug manufacturing fails in the face of the evolutionary approach.Evolved molecules are as effective as rationally designed drugs.The only difference is that we know nothing about how or why it works.We just know it passed all the tests.These invented drugs are beyond our comprehension, they are the product of "irrational design".

Pharmaceutical evolution allows researchers to be ignorant, while evolution itself is slowly getting smarter.Andrew Ellington, an evolutionary biochemist at Indiana State University, told Science that in an evolutionary system "it's about having the molecule tell you something about it, because it knows itself better than you do." Fertility drugs can benefit medicine.But drugs are not the same as software.We might be able to breed software, then put the system in its hands and let it breed on its own, into a realm that no one knows about.But can we put molecules on this evolutionary path to nowhere?

The answer is yes, but it would also be a daunting task.Tom Ray's electrical evolutionary machine focuses on processing heritable information, but ignores the organism; while molecular evolution focuses on the organism, but ignores heritable information.Pure information itself is difficult to eliminate, and there is no evolution without death.Muscle and blood are so helpful to evolution precisely because the organism provides a convenient way for information to die.Any system that can combine heritable information and perishable organisms has elements of an evolutionary system. San Diego biochemist Gerald Joyce studies the chemistry of early life.He proposes a simple method capable of incorporating the dual nature of evolution, information and organism, into a robust artificial evolutionary system: he recreates in a test tube what may have been the early stages of life on Earth - "RNA World ".

RNA is a very delicate molecular system.It is not the earliest life system, but life on earth will almost certainly become RNA life at a certain stage.Joyce said: "Everything in biology shows that the Earth 3.9 billion years ago was played by RNA." RNA has a unique advantage that is not shared by any other system known to us.It can be both the organism and the message at the same time—behavior and inner cause; messenger and message.A ribonucleic acid molecule should not only take on the responsibility of interacting with the world, but also complete the important task of continuing the world, at least passing the information to the next generation.Despite its heavy responsibility, ribonucleic acid is still an extremely compact system, from which open-ended artificial evolution can unfold.

The Scripps Research Institute is a sleek, modern laboratory by the sea near San Diego, California.Here, Gerald Joyce conducts his evolutionary experiments with a small group of graduate students and postdocs.At the bottom of the plastic test tube there is a little droplet, not as big as a thimble. This is his ribonucleic acid world.Dozens of these test tubes are placed in ice buckets, and when evolution is needed, they are heated to body temperature.Once warmed, ribonucleic acid can make a billion copies in an hour. "What we have," Joyce said, pointing to a small test tube, "is a massively parallel processor. One of the reasons I choose biological evolution over computer simulation is that on Earth, at least in the near future , no computer has yet given me 1015 parallel microprocessors." The droplet at the bottom of the test tube is roughly the same size as the intelligence on a computer chip.Joyce further elaborated: "In fact, our artificial system is even better than natural evolution, because there are not many natural systems that allow us to produce 1015 individuals in an hour."

In addition to the revolution in intelligence brought about by self-sustaining living systems, Joyce believed that evolution could bring commercial profits in the manufacture of chemicals and medicines.In his imagination, the molecular evolution system can run 24 hours a day, 365 days a year. "You give it a task and tell it not to leave the workshop until you figure out how to turn molecule A into molecule B." In one breath, Joyce named a long list of biotech companies that specialize in directed molecular evolution research (Gilead, Ixsys, Nexagen, Osiris, Selectide, and Darwin Molecular).His list also excludes established biotech companies such as Genentech, which is engaged in cutting-edge research not only in directed evolution but also in rational drug design.Darwin Molecules, whose main patent is held by complexity scientist Stuart Kaufman, has raised millions of dollars to harness evolution to design drugs.Nobel Prize-winning biochemist Manfred Eigen called directed evolution "the future of biotechnology."

However, is this true evolution?Is it the same evolution that gave us insulin, eyelashes and raccoons?Yes, this is evolution. “What we usually call evolution is Darwinian evolution,” Joyce told me, “but in another evolution, the selection pressure is determined by us, not by nature, so we call it directed evolution.” Directed evolution is another supervised learning, another way to traverse Borges' library, another kind of breeding.In directed evolution, selection is guided by breeders rather than naturally occurring.